Environmental Simulation Chambers: Application to Atmospheric Chemical Processes (NATO Science Series: IV: Earth and Environmental Sciences)

Environmental Simulation Chambers: Application to Atmospheric Chemical Processes

NATO Science Series A Series presenting the results of scientific meetings supported under the NATO Science Programme. The Series is published by IOS Press, Amsterdam, and Springer in conjunction with the NATO Scientific Affairs Division

Sub-Series I. II. III. IV.

Life and Behavioural Sciences Mathematics, Physics and Chemistry Computer and Systems Science Earth and Environmental Sciences

IOS Press Springer IOS Press Springer

The NATO Science Series continues the series of books published formerly as the NATO ASI Series. The NATO Science Programme offers support for collaboration in civil science between scientists of countries of the Euro-Atlantic Partnership Council. The types of scientific meeting generally supported are “Advanced Study Institutes” and “Advanced Research Workshops”, and the NATO Science Series collects together the results of these meetings. The meetings are co-organized bij scientists from NATO countries and scientists from NATO’s Partner countries –countries of the CIS and Central and Eastern Europe. Advanced Study Institutes are high-level tutorial courses offering in-depth study of latest advances in a field. Advanced Research Workshops are expert meetings aimed at critical assessment of a field, and identification of directions for future action.

As a consequence of the restructuring of the NATO Science Programme in 1999, the NATO Science Series was re-organized to the four sub-series noted above. Please consult the following web sites for information on previous volumes published in the Series. http://www.nato.int/science http://www.Springer.com http://www.iospress.nl http://www.wtv-books.de/nato-pco.htm

Series IV: Earth and Environmental Series – Vol. 62

Environmental Simulation Chambers: Application to Atmospheric Chemical Processes edited by

Ian Barnes Bergische University Wuppertal, Physical Chemistry Department, Germany and

Krzysztof J. Rudzinski Institute of Physical Chemistry of the PAS, Warsaw, Poland

Published in cooperation with NATO Public Diplomacy Division

Proceedings of the NATO Advanced Research Workshop on Environmental Simulation Chambers: Application to Atmospheric Chemical Processes Zakopane, Poland 1 - 4 October 2004

A C.I.P. Catalogue record for this book is available from the Library of Congress. ISBN-10 ISBN-13 ISBN-10 ISBN-13 ISBN-10 ISBN-13

1-4020-4231-0 (PB) 978-1-4020-4231-7 (PB) 1-4020-4230-2 (HB) 978-1-4020-4230-0 (HB) 1-4020-4232-9 (e-book) 978-1-4020-4232-4 (e-book)

Published by Springer, P.O. Box 17, 3300 AA Dordrecht, The Netherlands.

www.springer.com

Printed on acid-free paper

All Rights Reserved © 2006 Springer No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work Printed in the Netherlands.

CONTENTS

Preface

ix

Acknowledgements

xiii

Simulation Chamber Facilities Overview on the Development of Chambers for the Study of Atmospheric Chemical Processes Becker, K. H.

1

The UCR EPA Environmental Chamber Carter, W. P. L.

27

Investigations of Secondary Organic Aerosol in the UCR EPA Environmental Chamber Cocker II, D. R. and Song, C.

43

Field Measurement and Modelling Motivated Design of a Programme of Aerosol Chamber Experiments McFiggans, G.

49

Chamber Simulations of Cloud Chemistry: The AIDA Chamber Wagner, R., Bunz, H., Linke, C., Möhler, O., Naumann, K-H., Saathoff, H., Schnaiter, M., and Schurath, U.

67

New on-line Mass Spectrometer for Identification of Reaction Products in the Aqueous Phase: Application to the O H-oxidation of N-methyl-pyrrolidone under Atmospheric Conditions Poulain, L., Monod, A., and Wortham, H.

83

Dynamic Chamber System to Measure Gaseous Compounds Emissions and Atmospheric-Biospheric Interactions Aneja, V. P., Blunden, J., Claiborn, C. S., and Rogers, H. H.

97

Chemical Process Studies Chamber Studies on the Photolysis of Aldehydes Wenger, J. C.

111

Determination of Photolysis Frequencies for Selected Carbonyl Compounds in the EUPHORE Chamber Olariu, R-I., Duncianu, M., Arsene, C., and Wirtz, K.

121

v

vi

Contents

Remote Sensing of Glyoxal by Differential Optical Absorption Spectroscopy (DOAS): Advancements in Simulation Chamber and Field Experiments Volkamer, R., Barnes, I., Platt, U., Molina, L. T., and Molina, M. J.

129

Aromatic Hydrocarbon Oxidation: The Contribution of Chamber Oxidation Studies Bloss, C., Jenkin, M. E., Bloss, W. J., Rickard, A. R., and Pilling, M. J.

143

FT-IR Kinetic Study on the Gas-Phase Reactions of the OH Radical with a Series of Nitroaromatic Compounds Bejan, I., Barnes, I., Olariu, R., Becker, K. H. and Mocanu, R.

155

Atmospheric Fate of Unsaturated Ethers Mellouki, A.

163

Atmospheric Oxidation of the Chlorinated Solvents, 1,1,1-Trichloroethane, Trichloroethene and Tetrachloroethene Nolan, L., Guihur, A-L., Manning, M., and Sidebottom, H.

171

Simulation Chamber Study of the Oxidation of Acetic Acid by OH Radicals: Detection of Reaction Products by CW-CRDS in the Near-Infrared Range Crunaire, S., Fittschen, C., Lemoine, B., Tomas, A., and Coddeville, P.

181

Kinetics, Products and Mechanism of O(3P) Atom Reactions with Alkyl Iodides Barnes, I.

193

Kinetics of the Reaction between CF2O and CH3OH Burgos Paci, M. A. and Argüello, G. A.

207

New Kinetic and Spectroscopic Measurements in the CF3Ox + NOx System Chiappero, M. S., Malanca, F. E., Argüello, G. A., Nishida, S., Takahashi, K., Matsumi, Y., Hurley, M. D., and Wallington, T. J.

213

Kinetic Study of the Temperature Dependence of the OH Initiated Oxidation of Dimethyl Sulphide Albu, M., Barnes, I., and Mocanu, R.

223

Environmental Chamber Studies of Ozone Formation Potentials of Volatile Organic Compounds Carter, W. P. L.

231

Evaluation of the Detailed Tropospheric Chemical Mechanism, MCM v3, Using Environmental Chamber Data: Butane and Its Degradation Products Pinho, P. G., Pio, C. A., and Jenkin, M. E.

241

Studies on Nitrate-Affected SO 2 Oxidation and Their Perspectives Pasiuk-Bronikowska, W. and Bronikowski, T.

253

Contents

vii

Heterogeneous and Aqueous-Phase Transformations of Isoprene Rudzinski, K. J.

261

Investigation of Atmospheric Transformation of Diesel Emissions in the European Photoreactor (EUPHORE) Zielinska, B., Sagebiel, J., Stockwell, W., McDonald, J., Seagrave, J-C., Wiesen, P., and Wirtz, K.

279

Investigation of Real Car Exhaust in Environmental Simulation Chambers: Results from the INFORMATEX and DIFUSO Projects Wiesen, P.

285

The EUROCHAMP Integrated Infrastructure Initiative Wiesen, P.

295

Air Pollution Studies Survey on Atmospheric Chemistry Research in Some New EU Member States and Candidate Countries Batchvarova, E., Spassova, T., Valkov, N., and Iordanova, L.

301

New Measurements of NMVOC Concentrations in the City Air of Wuppertal, Germany: Input Data for Chemical Mass Balance Modelling Niedojadlo, A., Becker, K. H., Kurtenbach, R., and Wiesen, P.

341

Surface and Total Ozone Over Bulgaria Kolev, S. and Grigorieva, V.

351

Heavy Metals Pollution: An Everlasting Problem Mocanu, R., Cucu-Man, S., and Steinnes, E.

359

Atmospheric Wet Deposition Monitoring in Iasi, Romania Arsene, C., Mihalopoulos, N., Olariu, R-I., and Duncianu, M.

369

Problems of Air Quality in Tashkent City Tolkacheva, G. A.

379

Precipitation Quality in Different Zones of the Tashkent Region in Relation to Photo-Chemical Reactions Smirnova, T. and Tolkacheva, G.

393

Influence of Atmospheric Aerosol Contamination on the Regional Climate in Central Asia Chen, B. B. and Lelevkin, V. M.

403

Assessment of Air Pollution and Ecosystem Buffer Capacity in the Industrial Regions of Ukraine Kharytonov, M. M., Anisimova, L. B., Gritsan, N. P. and Babiy, A. P.

415

viii

Contents

Two Neutral Network Methods in Estimation of Air Pollution Time Series Cigizoglu, H. K., Alp, K., and Kömürcü, M.

421

List of Participants

433

Author Index

441

Subject Index

443

PREFACE Atmospheric pollution has many different detrimental impacts on air quality at urban, regional and global scales. Large volume photoreactors (often referred to as smog or simulation chambers) have been used very effectively to investigate and understand many varied aspects of atmospheric chemistry related to air pollution problems. Photochemical smog formation, which was first observed around 1945 in Los Angeles, is now a major environmental problem for all industrialised and densely populated regions of the world. Over the years many different modelling and experimental tools have been developed to analyse and simulate the complex chemical processes associated with tropspheric photooxidant formation. Work in environmental chambers has played a key role in the development of our understanding of the atmospheric chemistry associated with pollution problems on local, regional and global scales. Chamber observations have also been used in connection with environmental policy issues. In general they are used for validation of atmospheric chemical models, studies of chemical reaction mechanisms and as a direct means to test the possible impact of specific chemical compounds on air quality under simulated ambient conditions New large smog chamber installations have been recently developed in the US (Riverside, California), Europe (Jülich, Germany) and Japan, and a large number of smaller scale laboratory chambers are in operation around the world. Over the years there have been numerous new technical developments related to environmental chamber facilities such as the design of the chambers (e.g. outdoor versus indoor chambers), the techniques applied to control the physical and chemical conditions within them and the analytical facilities that are coupled to the chambers. Some chambers are now equipped with technology allowing measurement of highly reactive radicals such as OH HO2 and NO3, that play pivotal roles in atmospheric chemistry. In addition to the technical developments the scientific applications of chambers are being expanded, for example, the current understanding of secondary organic aerosol formation has been developed principally from environmental chamber studies and chambers have been used even more recently to investigate photolysis processes under ”real” atmospheric conditions. Such new scientific applications are often linked to new technical analytical and other technical developments and environmental chambers offer a perfect platform for testing environmental monitors. The many new scientific applications and technical developments of environmental chambers in the last years created a need for an Advanced Research Workshop where scientists who use chambers (experimentalists, analyticlists as well as modellers) could meet to exchange ideas, experiences and results and document advances in the application of chambers to research. With this background a NATO ARW entitled “Environmental Simulation Chambers: Application to Atmospheric Chemical Processes” was organised and held in the Antalowka Hotel (www.polskietatry.pl.) in Zakopane, Poland from Friday 1st October to Monday 4th October 2004. The major goal of the workshop was to review i) the various types of photoreactors currently in use and ii) the considerable diversity of environmental problems to which photoreactors have been used to successfully investigate. The aims were timely since in Europe within the EU 6th Framework programme a project entitled EUROCHAMP was initiated which had the aim of integrating large scale chamber facilities in Europe and also eventually US facilities. Another goal of the meeting was to promote a better awareness of the scientific community, in particular from NATO partner/Mediterranean dialogue countries and the EU ix

x

Preface

countries, of the availability of such facilities and their considerable potential for investigating problems related not just too atmospheric processes but chemical processes in general. It was particularly hoped that the workshop would attract young scientists from NATO partner countries where such facilities are not widely available but with assistance of the expertise from NATO countries basic experimental setups could easily be realized. At the time of the organisation of the workshop the European Union was expanded by 10 New Member States (Cyprus, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, Slovakia and Slovenia). Blugaria, Romania and Turkey are at the time of writing EU applicant countries. Presently, chamber technology is mainly located in western European countries and the US, therefore, in order to actively integrate scientists from eastern countries into the workshop programme a special section was organized with presentations from invited key speakers from several of these countries which showcased many of the acute environmental problems within these countries. A highlight of this section was overview of the research within Eastern European Countries by Dr. E. Batchvarova (Bulgaria) which was specially commissioned by Joint Research Center, Ispra, Italy. The workshop created a unique forum for scientists to present the new scientific applications of environmental chambers as well as the technical developments of these installations, including the advantages and drawbacks of the different types of smog chambers that are in use. Scientists from eastern European countries had the opportunity to learn firsthand about chamber techniques and dialogue with chamber users helped to identify potential applications of the techniques to their own environmental problems. This book contains contributions which give a broad overview of the different types of chamber facilities available and their major applications. The contents has been arranged along three thematic lines – presentation of simulation chamber facilities, chemical process studies, mostly utilising the chamber facilities, and regional studies, mostly indicating the needs for chamber use. Restrictions in the size of the workshop meant that not all the major chambers users could be invited to the meeting but we hope that sufficient reference has been made within the various contributions to these facilities that will allow easy referencing by the reader. (i) The section on Simulation Chambers Facilities begins with an historical overview of chamber development in USA and Europe, discussion of the advantages and disadvantages of using the chambers, the experimental techniques applied and gives exemplary results obtained from chambers in Bonn (now non-operational), Wuppertal and Valencia [Becker]. Detailed overviews are provided of smog chambers in Riverside [Carter; Cocker, aerosol studies], Bayreuth [Zetsch, investigation on POPs], and Karlsruhe [Wagner]. Two untypical yet interesting microchambers are presented, that are being used for studies of aqueous-phase oxidation of N-methyl pyrrolidone (an industrial solvent) [Monod] and for the measurement of fluxes of gaseous compounds between water and air or earth and air [Aneja]. (ii) The section on Chemical Process Studies carried out in the chambers contains a general discussion on the photolysis of atmospheric trace components [Wenger], as well as studies on the photolysis of small carbonyl compounds [Olariu], the photolysis of nitroaromatic compounds [Bejan] and the photolysis of chloral and dichloroacetyl chloride [Sidebottom]. Further contributions in this section describe oxidation processes investigated in the chambers and include: the OH initiated oxidation of benzene, toluene and p-xylene leading to formation of glyoxal, that was monitored by DOAS both in a chamber and in the

Preface

xi

field [Volkammer], an overview on the oxidation of aromatic hydrocarbons [Pilling], reactions of nitroaromatic compounds with OH that contribute to the formation of aerosols [Bejan], the reaction of acetic acid with OH radicals studied with the Continuous-Wave Cavity Ring Down Spectroscopy [Crunaire], and the oxidation of alkyl iodides by O(3P) atoms [Barnes]. Other contributions discuss VOC reactivity and ozone formation potential [Carter], the atmospheric fate of unsaturated ethers [Mellouki], reactions of oxygenated fluorocarbons with NOx and methanol [Arguello; Arguello], evaluation of the MCM v.3 against chamber data [Pinho], and investigations on vehicle ehaust emissions [Wiesen; Zielinska]. Two contributions report on aqueous and heterogeneous chemistry, namely the influence of nitrate ions on the mechanism of S(IV) autoxidation [Pasiuk-Bronikowska], and the heterogeneous and aqueous-phase reactions of isoprene [Rudzinski]. (iii) The section on Regional Studies refers to the geographical area stretching from Central and Eastern Europe to Southern Europe and to Central Asia. The contributing countries were Bulgaria, Germany, Uzbekistan, Kyrghistan, Romania, Turkey and Ukraine. The section starts with contribution which surveys the state (at the time of writing) of atmospheric chemistry research in the new EU member and candidate countries [Batchvarova]. The section covers several important topics. Air quality and pollution in cities is discussed in terms of the contribution of vehicle traffic and solvent use [Niedojadlo], the monitoring and analysis of wet deposition (pH and insoluble fraction) [Arsene], the influence of ozone, dust, nitrogen oxides, carbon oxide and sulphur dioxide, as well as their transformation [Smirnova; Tolkacheva], and the contamination of aerosols [Lelevkin]. Behaviour and effects of heavy metals in the atmosphere are discussed [Mocanu] as well as their impact on the ecosystems [Kharytonov]. The latter contribution also analyses the capacity of ecosystems to cope with air pollution. Other contributions discuss monitoring of ozone and collecting of reference data during a solar eclipse [Kolev], using neural network approach to properly fill the gaps in monitoring data [Cigizoglu], as well as lidar analysis of long range transport of aerosols and source identification [Lelevkin]. In closing, environmental simulation chambers continue to serve as indispensable tools in atmospheric chemistry research. Laboratory and modelling studies are being more frequently coupled to simulation chamber experiments for verification of ideas and for obtaining a deeper insight and understanding of chemical processes. Noticeably, present uses of chambers are not being restricted to the study of gas phase processes, the need for the study of aqueous-phase and heterogeneous studies within a common chamber forum has been recognised and methods are emerging for the study of multiphase systems. Environmental simulation chambers are located mainly in the USA and Western Europe. It would be good if the existing chambers could be made available to researchers outside these confines possibly through common projects. Possible initiatives to improve the global disparity in distribution with newly built chambers is also an alternative which could be explored. Some support for such actions exist within European projects such as ACCENT, EUROCHAMP and INTROP (ESF), which were briefly introduced at the workshop.

Ian Barnes and Krzysztof J. Rudzinski, June 2005

ACKNOWLEDGEMENTS The NATO Science Committee is very gratefully acknowledged for having faith in the venture and providing the initial basis financial support for the workshop. The additional support which was received from other sponsors is also very gratefully acknowledged. Without the support of all the sponsors it would not have been possible to invite and support the 55 participants at the meeting and make the workshop such an all-round success. Financial support of the workshop was by the following organizations/industries: x NATO Scientific and Environmental Affairs Division, Brussels x Institute for Environmental and Sustainability, JRC x European Science Foundation Programme INTROP x BMW - Bavarian Motor Company Munich

xiii

Overview on the Development of Chambers for the Study of Atmospheric Chemical Processes Karl H. Becker Bergische University Wuppertal, FB C, Physical Chemistry, Gauss Str. 20, 42119 Wuppertal, Germany Key Words: Atmospheric chemistry, Hydroxyl radicals, Ozone, Photoreactors, Photosmog

Introduction Photoreactors have been used for quite some time by scientists to study organic photochemistry (Bayes, Blacet, Calvert, Gunning, Hammond, Heicklen, Leighton, Okabe, Noyes, Steacy, and others mentioned in the text book by Calvert and Pitts, 1966). In industry, the photolytic initiation of product synthesis has been known since the beginning of the last century. A large chamber, known as the “Große Bonner Kugel” (Groth et al., 1972), was built in Bonn between 1966 and 1968; its conception was initiated by Groth and Harteck to study air glow reactions at the low pressures prevailing in the upper atmosphere. Afterwards, the chamber was also used to study reactions of tropospheric interest, but the knowledge of tropospheric chemistry at that time was very basic. Photolytic radical sources could not be applied in this reactor, and the radicals had to be introduced from the exterior using discharge flow techniques. Details of the chamber in Bonn will be discussed later in the article. Shortly after the Second World War, the Los Angeles photosmog became a problem which was investigated using relatively large chambers for the simulation of plant damage and health effects, e.g. eye irritation (Hagen-Smit, 1952). In the mid 70-ties, the Pitts group in Riverside (Finlayson-Pitts and Pitts, 1986; 2000) started to build an indoor chamber with the objective of exploring the processes which lead to photosmog formation, see Figure 1. Advances were slow because appropriate analytical techniques still had to be developed. Following this activity, Akimoto duplicated the chamber in Japan (Akimoto et al., 1979), while Atkinson in the Pitts group started to develop very successfully the procedures to measure OH, O3 and NO3 reaction rate constants, after the importance of OH radical reactions for tropospheric chemistry had been accepted. At the same time, others started to use Teflon bags for studying smog reactions under irradiation by natural sunlight, but their results were very much limited to the conditions prevailing in L.A.. Rather important work began at Ford Motor Company where in the late 60-ties Weinstock (1969) promoted the importance of OH reactions in the troposphere. In this laboratory Niki used a rather small photoreactor to develop the application of FTIR spectroscopy for quantitative investigation of atmospheric reactions in the laboratory (Niki et al., 1972; Wu et al., 1976; Niki et al., 1981). IR absorption spectroscopy was applied quite early to study chemical reactions of atmospheric interest (Stephens, 1958; Hanst, 1971), based mainly on mirror systems which allowed long path light absorption (White, 1942; 1976; Herriott et al., 1964, 1965). However, real progress and success came with the application of modern FTIR spectrometers by Niki et al. (1981) and further work in the Pitts’ group to quantitatively measure rate constants and products in photoreactors by long path FTIR absorption. 1 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 1–26. © 2006 Springer. Printed in the Netherlands.

2

Figure 1.

K . H. Becker

Indoor chamber from Riverside, already employing long path FTIR absorption spectroscopy (Source: Finlayson-Pitts and Pitts, 2000).

In the late 60-ties the understanding of atmospheric reactions took a large step forward when OH radicals were identified as the most important oxidizing agent in the troposphere. The discovery resulted in particular from the analysis of the CO budget (Heicklen et al., 1969; Weinstock, 1969, Stedman et al., 1970; Levy, 1971), though the oxidation of CO by OH had already been measured two years earlier (Greiner, 1967). The built-up of a radical chain was understood 10 years later when the rate constant of the fast reaction HO2 + NO o OH + NO2 was measured by several groups (Howard and Evenson, 1977; Leu, 1979; Howard, 1979; Glaschick-Schimpf et al., 1979; Hack et al., 1980; Howard, 1980; Thrush and Wilkinson, 1981) after Crutzen and Howard (1978) indicated the important role of this reaction in stratospheric ozone chemistry. However, it turned out that OH/HO2 radicals are also the key species in the tropospheric oxidation processes that determine the smog formation. In Europe, studies of tropospheric reactions were started in the 70-ties by a few laboratories, e.g. Becker and coworkers in Bonn and Cox and coworkers in Harwell. The technique of long path FTIR absorption mentioned above was introduced in the early 80-ties by Becker and co-workers in Wuppertal. They used a multiple reflection mirror system in a 420 l photoreactor which can be operated between -50 to 50°C to determine the OH reaction rate constants in combination with product analyses in the ppm range, Somewhat later a 6 m long quartz glass reactor of 1080 l volume was built which enabled measurements to be extended down to the ppbV level. The reaction chambers are illustrated in Figures 12 - 16 later in the text where some examples of selected experimental work carried out in the chambers are also given. More recently, other European laboratories have started to use indoor chambers up to 1000 l volume irradiated by black lamps (Baltensperger et al. in Villigen/Zürich, Carlier and Doussin in Paris, Hjorth et al. in Ispra, Herrmann et al. in Leipzig, Le Bras et al. in Orléans, Treacy et al. in Dublin, Wenger et al. in Cork). Tables 1 and 2 show a, not necessarily complete, list of larger indoor and outdoor reactors.

Overview on the Development of Chambers Tabl e 1.

Indoor chambers without light sources, or irradiated by black lamps or solar simulators.

Year of Laboratories Building 1968 1976 1976

3

Groth et al., Bonn Pitts et al., Riverside Niki et al., Dearborn

Description

Application

dark chamber 220 m3, stainless steel (high vacuum) 6 m3, evacuable, thermostated, FEP coated aluminium 150 l Pyrex

without light sources, for low pressure studies photooxidant and kinetic studies kinetic and mechanistic studies

1979

Akimoto et al., 6 m3 evacuable, thermostated, FEP coated aluminium Tsukuba

photooxidant studies

1980

Winer et al., Los Angeles

6 m3 evacuable, thermostated, FEP coated aluminium

photooxidant studies

1981

Becker et al., Wuppertal

420 l Duran glass, evacuable, thermostated -50 to +50 ºC

gas phase studies

1982

Joshi et al. (EPA), Research Triangle Park

440 glass reactor

photooxidant studies

1986

Becker et al., Wuppertal

quartz glass 1080 l, evacuable, thermostated 0 to +25 °C

1986

Evans et al., Australia

4 x 200 l FEP bags

gas phase and aerosol kinetic and mechanistic studies photooxidants studies

1988

Zetzsch et al., ca. 3000 l, Duran glass, Hannover, now thermostated –25 oC to ambient temperature Bayreuth

aerosol studies

1990

Schurath et al., 84 m3, thermostated -90 to +60 ºC, AIDA Karlsruhe

without light source, for aerosol studies

1996

Wahner et al., Jülich.

256 m3, FEP wall cover of a lab room

without light source, for NOY chemistry

1997

Carlier et al., Paris

977 l, glass

gas phase mechanistic studies

1998

Seinfeld et al., Pasadena Carter et al., Riverside

2 × 28 m3, 10 to 40 ºC

aerosol studies

FEP, double wall

low NOx studies

2000

K. H. Becker

4 Table 2.

Outdoor chambers irradiated by sunlight.

Year of Building

Laboratories

Volume, Wall Material

1976

Jeffries et al., Chapel Hill 25 m3, FEP

1981

Lonneman et al.,

15-40 l, FEP bags

1981

Fitz et al.,

40 m3, FEP

1983

Spicer et al.,

17.3 m3, FEP

1985

Kelly et al.,

450-2000 l, FEP bags

1985

Kamens et al., Chapel Hill

25 m3, FEP

1985

Seinfeld et al., Pasadena

65 m3, FEP

1995

Becker et al., Valencia

2 x 200 m3, FEP, EUPHORE

2000

Wahner et al., Jülich

280 m3, FEP, double wall, SAPHIR

In the meantime several large outdoor chambers have been built in the US with support from EPA to study the mechanisms of smog chemistry. One aim was to determine the ozone isopleths under US conditions. These chambers were mainly constructed from bags made of FEP, with volumes up to 25 m3. They brought about a better understanding of smog chemistry, but no results were obtained that could be generalized, because of narrow simulation conditions requested by the US-EPA. In Riverside, Carter and coworkers http://pah.cert.ucr.edu/~carter/bycarter.htm) developed a method to define the ozone ( formation potential of VOC's by determining maximum incremental reactivity (MIR) factors using smog chamber data and chemical modelling. A similar method was introduced by Jeffries in Chapel Hill (http://airsite.unc.edu/), who also used a smog chamber. Other groups injected engine exhausts directly into a smog chambers, and studied the formation of ozone. Collected at very different VOC/NOX ratios, the results from the US could not be generalized and applied to conditions in other countries. In addition, Atkinson and coworkers (http://www.ph.ucla.edu/scpcs/faculty/atkinson.html) refined their method to determine the OH reactivity from chamber measurements, using structure activity relationships to calculate rate constants of OH radical reactions with VOC’s. In the mid 80-ties, Seinfeld (http://www.che.caltech.edu/faculty/seinfeld_j/) in Pasadena, started to use a 65 m3 outdoor chamber made of FEP to study the aerosol formation from the oxidation of aromatic and biogenic hydrocarbons. In the mid 90-ties Becker and co-workers, together with Millán, built in Valencia/Spain the first European large outdoor chamber called EUPHORE (European Photoreactor). Actually, it consists of two chambers made of FEP, each of 200 m3 volume (Becker, 1996). This facility became a centre for European laboratories to carry out mechanistic and kinetic studies. Real exhaust gases from gasoline and Diesel engines were also investigated with respect to ozone formation. The EUPHORE chambers are equipped with all the necessary analytical instrumentation, also for in situ HO2 and OH measurements. In 2000, the group of Ehhalt in Jülich, now guided by Wahner, built a new double wall

Overview on the Development of Chambers

5

outdoor chamber called SAPHIR, of 280 m3 volume, see Figure 2. The double wall made of FEP allows studies of oxidation processes at low NOx concentrations (below 1 ppbV). The Ehhalt group did pioneering work in field measurements of OH and HO2 concentrations, so SAPHIR is fully equipped with the most advanced in situ radical measurement techniques. A smaller double wall indoor chamber (see book contribution) was recently built by Carter in Riverside, to study tropospheric oxidation processes at low NOX concentrations.

Figure 2 .

The double wall outdoor chamber SAPHIR in Jülich, Germany.

Two other chambers were built in Germany, at the same time, for the study of aerosol processes. Already in 1986, Zetzsch and coworkers built in Hannover a 3000 l Duran-glass indoor chamber, covered inside with FEP, and irradiated by solar simulators. This facility has recently been moved to Bayreuth. In 1990, Schurath and co-workers (see contribution in this book) built in Karlsruhe a 84 m3 stainless steel chamber called AIDA, which could be operated between -90 and +60 ºC. EUPHORE will be described later in more detail. Other groups also now operate medium sized chambers: Baltensperger in Zürich, Wenger in Cork, Le Bras in Orleans, Herrmann in Leipzig, Hjorth in Ispra and Doussin in Paris.

6

K. H. Becker

Large static reactors which can be operated at atmospheric pressure Advantages  The probability of wall effects relative to gas reactions decreases with the size R of a reactor, K ~ c • 1/R.  The large reactor diameter favours the application of long path absorption spectroscopy in the UV, VIS and IR. *  Sampling of gas is not limited, and allows the use of a large variety of analytical instruments. *  The large volume inside a reactor provides space for in situ analytical equipment,* e.g. cryo-trapping of radicals and LIF.  Application of the relative rate method enables the determination of rate constants not influenced by impurities, which is a problem in absolute rate determination.  The study of radical reactions with a reactant using the relative rate method is not influenced by radical chains which may develop during the reaction process.  A mixture of reactants can be introduced together with air or inert gases as tracers.  Studies of aerosol formation, and the chemical composition and physical properties of the particles become possible, provided the temperature gradient in the chamber can be controlled. *) This allows the analysis of reaction mechanisms by measuring intermediate products and operation in the ppbV range, i.e. at nearly ambient concentrations.

Disadvantages  With increasing R the volume of the reactor increases by R3 which creates room problems and increases the costs.  Low rate constants for radical reactions, below 10-14 cm3/(molecule • s), cannot be measured because the change of the reactant concentration is given by k×[radical] and [radical] is low, e. g. for OH in the order of 107 -108 radicals/cm3.  The photolytic radical source in a chamber might be less controllable (e.g. wall source of HONO). Special Problems  The light source of indoor chambers may deviate from the sunlight and create different photochemistry.  In the outdoor chamber, the solar spectrum varies with cloud cover and solar zenith angle and the distribution of the radiation is difficult to model. Radical Sources The radical source is a crucial point in using reaction chambers for photooxidation studies. At the successful dawn of chamber work, direct radical measurements were not technically possible. Production of OH radicals in systems like NOX + VOC + air + light occurred mainly by uncontrolled surface processes producing HONO which is photolysed.

Overview on the Development of Chambers

7

Photolytical OH sources in indoor chambers: x

VOC + NOX + air + hQ (source not controlled, probably includes HONO photolysis)

x

HONO + hQ o OH + NO

x

O3 + H2O + hQ o 2 OH + O2

x

RONO + NO + air + hQ o ....... OH RONO + hQ o RO + NO o R'CHO + HO2 RO + O2 HO2 + NO o OH + NO2

x

H2O2 + hQ o 2 OH

x

HCHO + NO + air + hQ o ..... 2 OH + air HCHO + hQ o H + HCO H + O2 + M o HO2 + M HCO + O2 o HO2 + CO HO2 + NO o OH + NO2 Dark OH sources used in chambers: x HO2NO2 + NO + air o OH + 2 NO2 HO2NO2 + M o HO2 + NO2 + M HO2 + NO o OH + NO2 x

olefins + O3 o .....OH......

x

N2H4 + O3 o ..... OH ......

NO3 radical source used in chambers: x N2O5 + M o NO3 + NO2 + M

x NO2 + O3 o NO3 + O2 RO2-radical rource used in chambers: x Cl2 + RH + O2 + hQ + M o 2 RO2 + 2 HCl + M The above lists shows the reactions which have been used to produce OH radicals under controllable conditions by irradiation with an appropriate light source or in the dark. For the study of NO3 radical reactions, either N2O5 has to be introduced into the chamber or the studies have to be carried in the presence of NO2 + O3. For RO2 radical studies, H atom abstraction by photolytically generated Cl atoms in the presence of O2 was employed. For indoor chambers, a variety of photolysis lamps have been used, either intensive Xe high pressure arc or fluorescent lamps with different spectral transmittances. The reactor walls for indoor chambers had to be made from Duran/Pyrex glass or quartz glass in order to ensure transmittance of photolytically active light for initiation of the radical photooxidation (see Figure 3) or the lamps had to be installed inside the reactor. For simulation studies the radical source should be as similar as possible to the outdoor situation, however, in reality this is never fully achieved. Figure 4 gives spectral thresholds for the production of photolytic radical sources in the troposphere, and Figure 5 shows the diurnal variation of the production rates of the different OH radical sources based on field measurements in Germany 1998 within the BERLIOZ campaign. This shows that different OH sources even during daylight

K. H. Becker

8

are active in the troposphere near the surface. Apart from simulation studies photoreactors of all sizes have been successfully used for a long time to measure reaction rate constants of radicals by employing the relative rate technique.

Figure 3 .

Absorptions spectra of NO2, HONO and CH3ONO.

Figure 4 .

Spectral thresholds of photolytic radical sources in the troposphere and the spectral distribution of sunlight a) above the ozone layer, b) near the surface.

Overview on the Development of Chambers

9

Figure 5 . Comparison of the four dominant atmospheric OH radical sources (photolysis of O3, HONO, HCHO, and ozonolysis of VOC’s) for 20th and 21st of July 1998. The difference in the OH production in the early morning hours can be seen easily. After 12:00 on 20th July and 8:00 on 21st July, respectively, the modelled J(HCHO) values were used, because no measurements were available for this time period. The dotted line shows the calculated P(OH) from nitrous acid photolysis during daytime. The data were obtained during the BERLIOZ campaign at the Pabstthum site, northwest from Berlin (Platt et al., 2002).

Relative rate technique The relative rate technique has a few advantages over other methods, i.e. no absolute concentrations have to be measured, impurities do not generally disturb the measurements and the experiments can be carried out in the presence of several reaction partners (Barnes et al., 1982). Also the initiation of radical chains during the reaction process does not distrurb the measurements. However, the rate constant of the reference reaction, k2, has to be known. The following system illustrates the kinetics of the relative rate technique.

10

K. H. Becker X1 + Y Products, k1 (cm3 s-1) X2 + Y Products, k2 X1 and X2: stable species, Y: radical

Reactions:

d [X1] / dt d ln [X1] / dt d [X2] / dt d ln [X2] / dt

= = = =

- k1 [X1] [Y] - k1 [Y] - k2 [X2] [Y] - k2 [Y]

(1/k1) ³d ln [X1]t = - ³[Y]t dt t

t

(1/k2) ³d ln [X2]t = - ³[Y]t dt t

t

(1/k1) ³d ln [X1]t = (1/k2) ³d ln [X2]t t

t

k1 / k2 = ln ([X1]t / [X1]o) / ln ([X2]t / [X2]o)

Selected experimental results from chamber studies "Große Bonner Kugel" The chamber was built between 1966-68 by W. Groth upon the initiative of P. Harteck. The facility was operated by K. H. Becker, E. Fink, D. Kley and U. Schurath (Groth et al., 1972). Properties of the chamber: volume: 221 m3 surface: 177 m2 inner diameter: 7.5 m heatable: Tmax = 350 ºC with 233 KW power for reaching a vacuum of 10-9 Torr. cooling of baffles between pumps and chamber with liquid hydrogen material: stainless steel, 10 mm wall thickness pumping speed: 240 000 l/s with 8 diffusion pumps lowest pressure: 10-9 Torr Major purpose: Studies of O and N atom reactions of aeronomic interest in the upper atmosphere in the pressure range of 0.1-100 mTorr. Major achievements: New results on the recombination reactions of NO + O + (M) (Becker et al., 1972c; 1973), N + N + (M) (Becker et al., 1969; 1971a; 1972b), reactions of O2(1'g) and O2(16+g) with other trace gases (Becker et al., 1971b; 1972a; Penzhorn et al., 1974), reactions of O3 with olefins (Becker et al., 1974a), detection of an energy transfer reaction from O2(1'g) to the HO2 radical and identification of a low lying excited state of HO2 (Becker et al., 1974b; 1975; 1978a). The studies on NO + O and N + N included the quantification of radiative recombination by inverse predissociation which could be clearly

Overview on the Development of Chambers

11

identified in the low pressure regime. Unfortunately, at that time dark OH radical sources and the importance of OH reactions were not known. Figures 6 - 8 show the facility installed beside the Institute of Physical Chemistry, Bonn University. The facility was dismantled few years ago.

Figure 6 .

The housing of the “Große Bonner Kugel”.

Figure 7. Cross-section through the spherical reaction chamber.

12

K . H. Becker

Figure 8 . The pipe system by which the chamber could be heated up to 300 °C (top), enormous pumping capacity needed to reach the vacuum of 10-9 Torr (middle), and the platform at which the experiments were prepared and carried out (bottom).

Overview on the Development of Chambers

13

Studies on the air glow reaction NO + O + M o NO2 + M + hQ The controversial discussion surrounding this reaction as to the relative contributions of a two-body or three-body mechanism could only be settled by experiments carried out at relatively low total pressure, because nearly all previous studies were performed in flow tubes at about 1 Torr or in flow tubes at lower pressures, with significant wall influences. It was expected that the large reaction chamber offered the best possibility to study the reaction at low pressures. The O atoms were generated by flowing O2 through a discharge before introducing the atoms into the chamber. This could be done from the side or even from the centre of the sphere by means of inserted Teflon tubes. After switching off the discharge, the atoms decayed rapidly by wall reactions. However, the large volume provided about 100 s time during the decay, in which the air glow could be measured during stepwise increase of the total pressure. Figure 9 shows an intensity-time profile of the air glow during the stepwise increase of the total pressure.

Figure 9 . The intensity-time profile of the air glow reaction NO + O + M during the decay of the O atoms and the stepwise increase of the total pressure, M = O2, from 0.1 to 0.8 mTorr (upper part) and from 0.1 to 24.5 mTorr (lower part). As can be seen from Figure 11, below 1 mTorr the air glow rate constant follows a two-body radiative recombination with a rate constant of 4.2 × 10-18 cm3 molecule-1 s-1

14

K . H. Becker

Figure 10. Spectral distribution of the air glow emission at different total pressures (M = O2), absolutely calibrated against the distribution at 1 Torr measured in a flow tube (Fontijn et al., 1964).

Figure 11. The air glow intensity obtained by the integral over the spectral distribution of the chemiluminescence shown in Figure 10, which is plotted as apparent two-body rate constant kM against total pressure (M = O2).

Overview on the Development of Chambers

15

whereas above 1 Torr the “high pressure” limit with an apparent two-body rate constant of 6.4 × 10-17 cm3 molecule-1 s-1 has been reached where three-body recombination by M and quenching by M is balanced. From the slope of the plot of kM vs. total pressure a three-body rate constant of 7 × 10-32 cm6 molecule-1 s-1 was determined. As a consequence, the air glowreaction proceeds by a two-body as well as by a three-body reaction. When the air glow rate constant is compared with the total recombination rate of NO + O + M, which also was measured in the chamber, it can be concluded that the air glow rate constant represents within the error limits the total recombination rate constant. Further details are given in the publications Becker et al. (1972c) and (1973). Experiments from Wuppertal Several photoreactors between 420 to 1080 l volumes are in operation in the Physical Chemistry Department of the Bergische University Wuppertal. A thermostated photoreactor of 420 l volume is shown schematically in Figure 12. The reactor consists of an inner Duran glass tube, o.d 0.6 m and length 1.5 m which is concentrically inserted in a wider Duran glass tube of the same length. On both sides the tubes are closed by aluminium flanges covered with Teflon on the inner side. Between the tubes a cooling fluid circulates, and inside the fluid space black lamps/fluorescent lamps in protective Duran covers are installed. The covers are flushed with dry purified air. Figure 13 shows that the whole reactor including the cooling systems is inserted into a rigid optical bench which is part of a White multiple path mirror system coupled to a FTIR spectrometer. In order to minimise light losses of the IR beam, the White mirrors are installed inside the reactor, but can be adjusted to optimal path length from outside. The flanges contain several ports, such as gas inlet/outlet ports and syringe port for taking GC samples, as well as windows for UV/Visible absorption spectroscopy. The optimal path length of several hundred meters enables quantitative concentration measurements in the

Figure 12 . Schematic view of the 420 l reactor in Wuppertal, operating in the temperature range of ± 50 oC.

K . H. Becker

16

ppmV range, while gases with low IR absorption are measured by GC. The spectroscopic measurements require careful recording of reference spectra. The reactor is evacuable down to 1 mTorr, and is operated mostly at atmospheric pressure, filled either with synthetic air or with rare gases. The temperature can be set between –50 to +50 °C. A first version of a

Figure 13.

The suspended double wall reactor of 420 l volume.

Figure 14 .

View of the thermostated and non-thermostated reactors mounted in series.

Overview on the Development of Chambers

17

photoreactor of the same size was built with the cooling system, and now it is installed in line with the thermostated reactor, Figure 14.

Figure 15 . View of the 6 m long quartz glass reactor of 1080 l volume.

Figure 16 .

Schematic drawing of the 1080 l reactor setup.

K . H. Becker

18

Figure 15 and 16 show the 1100 l quartz glass reactor consisting of two combined tubes with a combined length of 6.2 m length with an o.d. of 0.47 m. The end flanges contain all the necessary ports for gas inlets, gas sampling and the White mirrors. The path length for FTIR absorption was extended to approximately 500 meters, to achieve a sensitivity for product analysis reaching the low ppbV range. Examples of experimental work using the chambers (i)

Measurement of pressure dependence of the reaction rate constant of C2H4 + OH in the presence of air and argon (Klein et al., 1984),

(ii)

Measurement of the temperature and pressure dependence of the reaction rate constant HO2NO2 + OH (Barnes et al., 1981; 1986),

(iii)

Measurement of thermal decomposition rates of peroxynitrates, RO2NO2 (Zabel et al., 1986; Kirchner et al., 1990; Becker et al., 1993 and Zabel, 1995) as a function of temperature.

(i) Measurements of the pressure dependence of the reaction rate constant k1 for C2H4 + OH +M o products with M = air and Ar A dark source of OH radicals from the thermal decomposition of HO2NO2 in the presence of NO was employed, and the relative rate method was used with reference reactions, n-hexane + OH and n-butane + OH reactions, for which no pressure dependence is

Figure 17. Dependence of C2H4 + OH +M on total pressure at 295 K for M = synthetic air (M = Ar not shown). The solid line through the data points is the calculated falloff curve. The error bar represents the average of 2V errors for all individual data points.

Overview on the Development of Chambers

19

known. Figure 17 shows the experimental results and, together with Figure 18, demonstrates the high quality of the measurements.

Figure 18 . Comparison of literature k1 values near room temperature with the calculated falloff curves from this work (solid line M = Ar, broken line M = synthetic air. The data points are presented without error limits to simplify the comparison.

(ii) Studies on the temperature and pressure dependence of the reaction rate constant HO2NO2 + OH. The next example describes the measurements of the reaction rate constant of peroxynitric acid (HO2NO2) with OH radicals as a function of pressure (He) and temperature, details of which can be found in the literature (Barnes et al., 1981; 1986). It is impressive that for a rather complex chemical system no absolute concentration measurements were needed. A relative rate method with reference to the known reaction rate constant kn-butane + OH was used. HO2NO2 was prepared outside the reactor and introduced together with NO, He and the reference substance (n-butane) into the reaction chamber. The OH radicals steadily produced from the thermal decomposition of HO2NO2, caused a decay in n-butane which could be followed by GC analysis, stage I in Figure 19. The decay of HO2NO2 measured by FTIR absorption is due to thermal decomposition, possible wall losses and the HO2NO2 + OH reaction, also shown in Figure 19, stage I. After stage I, the n-butane concentration was rapidly increased, which resulted in a change of the OH concentration because of the faster radical consumption by n-butane. Consequently, the decay of HO2NO2 became slower because of the lower OH concentration. From the different slopes of HO2NO2 and n-butane in stage I and stage II, all measured relatively during the time steps 't in both stages, the

20

K . H. Becker

following equation was derived from which the reaction rate constant kHO2NO2 evaluated:

+ OH

was

kHO2NO2+OH/kn-butane+OH = {('ln[HO2NO2]/'t)stage I  ('ln[HO2NO2]/'t)stage II}/ {('ln[n-butane]/'t)stage I  ('ln[n-butane]/'t)stage II}

Figure 19. Concentration-time profiles for X = HO2NO2 (o) and X = n-butane (); reaction conditions: T = 278 K, ptotal = 133 mbar, M = He, [NO]o = 2.4 x 1015 molecules/cm3 , [NO2 ]o = 1.0 x 1014 molecules/cm3. (iii) The thermal decomposition rates of peroxynitrates, RO2NO2, as a function of temperature and pressure. The last example deals with measurements of the thermal decay of peroxynitrates, RO2NO2, as a function of pressure and temperature. Details can be found in the publications Zabel et al. (1986), Kirchner et al. (1990), Becker et al. (1993) and Zabel (1995). In the experiments, the peroxynitrates were produced in the presence of O2 and NO2 from R radicals generated by H atom abstraction by Cl atoms from different hydrocarbons (see the reaction scheme in Figure 20).

Overview on the Development of Chambers

21

Alkane + Cl2 +O2 + M + hQ

ĺ

RO2 +HCl +M

RO2 + NO

ĺ

RO + NO2 k1

ĺ

RONO2

RO2 + NO2 + M

ĺ

RO2NO2 + M

k2

RO2NO2 + M

ĺ

RO2 + NO2 + M

k3

(+M)

k3(effective) = k3 x {1 + (k2/k1)x[NO2]/[NO]}-1 k3(T, [M]) for [NO] >> [NO2] and (k2/k1) from variation of [NO]

Figure 20. Production and decomposition of peroxynitrates in the presence of Cl2, NO, NO2 and O2 during the photolysis of Cl2.

Figure 21. Decay of RO2NO2, here HO2NO2, at different NO concentrations.

According to the equation in Figure 20, by increasing the concentration of NO, the parameter k3 can be determined as a function of temperature and total pressure. This can be seen in Figure 21 where k3 is derived from the slope of ln(ct/co) a t at higher NO concentration. Table 3 shows the results for different peroxynitrates. It can be seen that a

22

K . H. Becker

carbonyl group beside -O2 stabilizes the peroxynitrate as for peroxyacetylnitrat, CH3C(O)O2NO2, (PAN), in particular at lower temperatures. Table 3 .

Rate constants, Arrhenius parameters and thermal lifetimes of several peroxynitrates (experiments and evaluation are described in Zabel et al. (1986), Kirchner et al. (1990), Becker et al. (1993) and Zabel (1995).

R in RO2NO2 C4H9 C6H13 C8H17 C2H5 CH3C(O)CH2 CH3 C6H5CH2 HOCH2 CH3OCH2 CH3C(O)OCH(CH3) C6H5OCH2 H CH3OC(O) CH3C(O) C6H5C(O) *

k3 (298K)*

Ea

Thermal lifetime, K-1

A

-1

kJ/mol

10 s

25 °C

0 °C

- 20 °C

6.5 5.8 3.7 3.7 2.6 1.65 1.5 1.0 0.44 0.24 0.23 0.083 0.00084 0.00040 0.00031

86.2 86.2 86.2 86.2 85.6 83.8 90.3 85.3 92.2 94.9 95.0 89.2 107.0 112.9 116.4

0.83 0.75 0.48 0.47 0.26 0.08 1.0 0.09 0.64 1 1 0.036 0.48 2.5 7.9

0.15 s 0.17 s 0.27 s 0.27 s 0.38 s 0.61 s 0.67 s 1.00 s 2.27 s 4.17 s 4.35 s 12.05 s 19.8 min 41.7 min 53.8 min

3.7 s 4.1 s 6.5 s 6.6 s 9.2 s 13.5 s 19 s 23.3 s 1.1 min 2.4 min 2.5 min 5.4 min 17.2 h 1.8 days 2.7 days

1.3 min 1.4 min 2.2 min 2.2 min 3.0 min 4.2 min 7.3 min 7.7 min 28.3 min 1.1 h 1.1 h 2.0 h 30 days 94 days 157 days

s

16 -1

k3 (T) = A×exp(-Ea/RT)

EUPHORE: the EUropean PHOtoREactor in Valencia In 1995, at the CEAM Foundation in Valencia, the European photoreactor came into operation. The facility was built by the group of Becker from Wuppertal, and is hosted and operated by CEAM under the direction of Millían and Wirtz. Figures 22 and 23 show photos of the facility, which consists of two half-sphere outdoor chambers made from TEDLAR. Each chamber has a volume of 200 m3. Movable rigid covers protect both chambers against heavy storms, and are used to cut off the solar irradiation during experiments. The chambers are fully equipped with state-of-the-art analytical instrumentation. Besides more conventional methods, FTIR long path absortion and DOAS are employed. In one chamber OH and HO2 radicals can be measured in situ by the LASER fluorescence technique. Regular EUPHORE reports are published which can be obtained from Ian Barnes ([email protected]) or Klaus Wirtz ([email protected]).

Overview on the Development of Chambers

Figure 22 . One of the two EUPHORE chambers at CEAM/Valencia.

Figure 23 . Building hosting two EUPHORE chambers, one closed by a metal cover.

23

24

K . H. Becker

Figure 24 . In situ measurements of OH radicals (line) during the oxidation of toluene, for comparison, the points represent OH concentrations that were calculated from the decay rate of toluene (Pilling, M. J. et al., 2003). An example of the good agreement which has been obtained in the EUPHORE chamber between measured OH radical profiles using the FAGE technique (Pilling, M. J. et al., 2003) and the OH radical concentration in the chemical system calculated from the decay of toluene is shown in Figure 24. Outlook Numerous results from kinetic and mechanistic studies using photo-reactors have been obtained which have contributed significantly to an increase in our knowledge on atmospheric chemistry, in particularly in the gas phase. Some open questions still remain, e.g. the mechanisms of the photooxidation of aromatic hydrocarbons and the process of gas-toparticle conversion. More recently also aerosol studies are being carried out in chambers. Some very recent work has led now to a much better understanding of the photochemical wall effects in chambers which produce OH radicals (Rohrer et al., 2005). The state of chemical knowledge has, however, so much advanced that it is quite difficult to propose further innovative work and the interest on additional applications of new results for improving tropospheric models is decreasing. On the other side, interest in exposure experiments using advanced simulation facilities and analytical techniques to study impacts of air pollutants on human health, plant damages and material destruction is on the increase. This could be viewed perhaps as a return to the 50-ties (Hagen-Smit, 1952), however, compared to then present-day approaches are at a much higher level of understanding of the chemical systems and use techniques which enable in situ radical measurements and the monitoring of trace gases in the pptV range. At the time of writing a new EU infrastructure project EUROCHAMP (co-ordinator P. Wiesen < [email protected] >) has as its main goal the integration of all European chamber facilities to improve studies on atmospheric chemical systems and also to open simulation facilities for impact studies.

Overview on the Development of Chambers

25

References Akimoto, H., M. Hoshino, G. Inoue, F. Sakamaki, N. Washida and M. Okuda; Environ. Sci. Technol. 13 (1979) 471. Barnes, I., V. Bastian, K.H. Becker, E.H. Fink and F. Zabel; Atmos. Environ. 16 (1982) 545. Barnes, I., V. Bastian, K. H. Becker, E. H. Fink and F. Zabel; Chem. Phys. Lett. 83 (1981) 459. Barnes, I., V. Bastian, K.H. Becker, E.H. Fink and F. Zabel; Chem. Phys. Lett. 123 (1986) 28. Becker, K.H., W. Groth and D. Kley; Z. Naturforschung 24a (1969) 1840. Becker, K.H., A. Elzer, W. Groth and D. Kley; Z. Naturforschung 26a (1971a) 929. Becker, K.H., W. Groth and U. Schurath; Chem. Phys. Lett. 8 (1971b) 259. Becker, K.H., W. Groth and U. Schurath; Chem. Phys. Lett. 14 (1972a) 489. Becker, K.H. E. H. Fink, W. Groth, W. Jud and D. Kley; Faraday Disc., Chem. Soc. 53 (1972b) 35. Becker, K.H., W. Groth and D. Thran; Chem. Phys. Lett. 15 (1972c) 215. Becker, K.H., W. Groth and D. Thran; 14th Symp. (International) on Combustion, Pittsburgh (1973) 353. Becker, K.H., U. Schurath and H. Seitz; Int. J. Chem. Kinet. 6 (1974a) 725. Becker, K.H., E. H. Fink, P. Langen and U. Schurath; J. Chem. Phys. 60 (1974b) 4623. Becker, K.H., E. H. Fink, P. Langen and U. Schurath; 15th Symposium (International) on Combustion, The Combustion Institute, Pittsburgh (1975) 961. Becker, K.H., E. H. Fink, A. Leiss and U. Schurath; Chem. Phys. Lett. 54 (1978a) 191. Becker, K.H., J. Löbel und U. Schurath; Staub-Reinhaltung Luft 38 (1978b) 278. Becker, K.H., F. Kirchner and F. Zabel; in: K.H. Becker and H.Niki (eds), The Tropospheric Chemistry of Ozone in Polar Regions. NATO ASI Series, Vol. 17, Springer Verlag, Berlin Heidelberg (1993) 351. Bridier, I., F. Caralp, H. Loirat, R. Lesclaux, B. Veyret, K. H. Becker, A. Reimer and F. Zabel; J. Phys. Chem. 95 (1991) 3594. Calvert, J.G. and J.N. Pitts, Jr.; Photochemistry, John Wiley & Sons, New York (1966). Crutzen, P.J. and C.J. Howard; Pure Appl. Geophys. 116 (1978) 497. Finlayson-Pitts, B.J. and J.N. Pitts, Jr.; Atmospheric chemistry, John Wiley & Sons, New York (1986). Finlayson-Pitts, B.J. and J.N. Pitts, Jr.; Chemistry of the upper and lower atmosphere, Academic Press, New York (2000). Fontijn, A., C.B. Meyer and H.I. Schiff; J. Chem. Phys. 40 (1964) 64. George, C., R. S. Strekowski, J. Kleffmann, K. Stemmler and M. Ammann; Faraday Discuss. 130 (2005) 1-16. Glaschick-Schimpf, I.,A. Leiss, P.B. Monkhouse, U. Schurath, K.H. Becker and E.H. Fink, Chem. Phys. Lett. 67 (1979) 318. Greiner, N. R.; J. Chem. Phys. 46 (1967) 2795. Groth, W., K. H. Becker, G.H. Comsa, A. Elzer, E. Fink, W. Jud, D. Kley, U. Schurath and D. Thran; Naturwissenschaften 59 (1972) 379. Haagen-Smit; Ind. Eng. Chem. 44 (1952) 1342. Hack, W., A. W. Preuss, F. Temps, H.Gg. Wagner and K. Hoyermann; Int. J. Chem. Kinet. 12 (1980) 851. Hanst, P. L.; Adv. Environ. Sci. Technol. 2 (1971) 91. Heicklen, J., K. Westberg and N. Cohen; Report, Center for Air Environmental Studies, (1969) 115-169. Herriott, D., H. Kogelnik and R. Kompfner; Appl. Opt. 3 (1964) 523. Herriott, D., and H.J. Schulte; Appl. Opt. 4 (1965) 883.

26

K . H. Becker

Howard, C.J., and K.M. Evenson; Geophys. Res. Lett. 4 (1977) 437. Howard, C.J.; J. Chem. Phys. 71 (1979) 2352. Howard, C.J.; J. Am. Chem. Soc. 102 (1980) 6937. Klein, Th., I. Barnes, K.H. Becker, E.H. Fink and F. Zabel; J. Phys. Chem. 88 (1984) 5020. Leu, M.-T.; J. Chem. Phys. 70 (1979) 1662. Levy, H., Science 173 (1971) 141. Niki, H., E.E. Daby and B. Weinstock; Adv. Chem. Ser., Washington, 113 (1972) 16. Niki, H., P.D. Maker, C.M. Savage and L. P. Beitenbach; Chem. Phys. Lett. 80 (1981) 499. Penzhorn, R.D., H. Güsten, U. Schurath and K.H. Becker; Environ. Sci. Technol. 8 (1974) 907. Pilling, M. J. et al. (2003). Effects of the oXidation of Aromatic Compounds in the Troposphere (EXACT, EVK2-CT-1999-00053), Final Report., http://www.chem.leeds.ac.uk/exact Platt, U., B. Alicke, R. Dubois, A. Geyer, A. Hofzumahaus, F. Holland, M. Martinez, D. Mihelcic, T. Klüpfel, B. Lohrmann, W. Pätz, D. Perner, F. Rohrer, J. Schäfer and J. Stutz; J. Atmos. Chem. 42 (2002) 359. Rohrer, F., B. Bohn, T. Brauers, D. Brüning, F.-J. Johnen, A. Wagner, J. Kleffmann, Atmos.Chem. Phys. Discus. 4 (2004) 7881. Stedman, D.H., E.D. Morris, Jr., E.E. Daby, H. Niki and B. Weinstock; 160th National Meeting of Am. Chem. Soc., Chicago (1970). Stephens, E.R.; Soc. Appl. Spectrosc. 12 (1958) 80. Thrush, B.A. and J.P.T. Wilkinson; Chem. Phys. Lett. 81 (1981) 1. Weinstock, B.; Science 166 (1969) 224. White, J.U.; J. Opt. Soc. Am. 32 (1942) 285. White, J.U.; J. Opt. Soc. Am. 66 (1976) 411. Wu, C.H., S.M. Japar and H. Niki; J. Environ. Sci. Health, Environ. Sci. Eng. A11 (1976) 191. Zabel, F.; Z. Physi. Chem. 188 (1995) 119.

The UCR EPA Environmental Chamber William P. L. Carter College of Engineering Center for Environmental Research and Technology, University of California, Riverside, California, 92521, USA Key Words: Environmental chambers, Atmospheric chemical mechanisms, Ozone, Volatile organic compounds, Oxides of nitrogen

Abstract The UCR EPA chamber is a new large indoor environmental chamber constructed at the University of California at Riverside (UCR) under United States EPA funding for the purpose of evaluating gas-phase and secondary aerosol mechanisms for ground-level air pollution. The major characteristics of this chamber, the results of its initial characterization for gas-phase mechanism evaluation, and examples of initial gas-phase mechanism evaluation experiments, are described. It is concluded that the chamber has lower or at most comparable background effects than other chambers previously used for mechanism evaluation, and can provide useful mechanism evaluation data at NOx levels as low as 2 ppb. Future research directions to utilize the capabilities of this chamber are discussed. Background Chemical mechanisms are critical components of airshed models used for predictions of secondary pollutants such as ground-level ozone or secondary organic aerosol. Because many of the chemical reactions are incompletely understood, these mechanisms cannot be relied upon to give accurate predictions of impacts on emissions on air quality until they have been shown to give accurate predictions under realistic but controlled conditions. The most reliable way to test this is to compare their predictions against results of well-characterized environmental chamber experiments that simulate the range of conditions in the atmosphere. If a model cannot accurately predict results of such experiments, it cannot be expected to reliably predict effects of proposed control strategies on ambient air quality. For this reason, environmental chambers are essential to developing predictive mechanism for compounds for which basic mechanistic information are insufficient (as is the cased for aromatics), testing approximations and estimates necessary for modelling the reactions of almost all VOCs under simulated atmospheric conditions, and testing entire mechanisms under varied conditions. However, results of environmental chamber experiments are only useful for testing mechanisms if the conditions of the experiments are sufficiently well characterized so that the characterization uncertainties are less than the uncertainties of the mechanisms being tested. As discussed by Dodge (2000), the chamber data base used to develop and evaluate current mechanisms had a number of limitations and data gaps that could affect their accuracy. Uncertainties exist concerning characterization of chamber conditions that could cause compensating errors in the gas-phase mechanism (Carter and Lurmann, 1990, 1991; Jeffries et al, 1992). Most chamber experiments lack measurement data for important species, limiting the level of detail to which the mechanisms can be evaluated, and the types of air quality impact predictions that can be assessed. Furthermore, because of chamber effects and because of inadequate analytical equipment employed, the current environmental chamber data base is not suitable for evaluating chemical mechanisms under the lower NOx conditions found in 27 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 27–41. © 2006 Springer. Printed in the Netherlands.

W. P. L. Carter

28

rural and urban areas with lower pollutant burdens. Because of this, one cannot necessarily be assured that models developed to simulate urban source areas with high NOx conditions will satisfactorily simulate downwind or cleaner environments where NOx is low. To address the need for improved an improved environmental chamber facility to evaluate mechanism for O3 and PM formation, the College of Engineering, Center for Environmental Research and Technology (CE-CERT) at the University of California at Riverside (UCR) has undertaken a program to develop a “Next Generation” environmental chamber facility for chemical mechanism evaluation and VOC reactivity assessment. The objectives are to develop the environmental chamber facility needed for evaluating gas-phase and gas-to-particle atmospheric reaction mechanisms, for determining secondary aerosol yields, and for measuring VOC reaction products and radical and NOx indicator species under more realistic and varied environmental conditions than previously has been possible. This project resulted in the construction of new “UCR EPA” chamber, which became fully operational in early 2003. The design features and characteristics of this chamber, the results of its initial characterization, and completed and ongoing projects in this chamber are summarized briefly below. Facility Description The indoor facility comprises a 20’x20’x40’ thermally insulated enclosure that is continually flushed with purified air at a rate of 1000 L min-1 and is located on the second floor of a laboratory building specifically designed to house it. Located directly under the enclosure on the first floor is an array of gas-phase continuous and semi-continuous gas-phase monitors. Within the enclosure are two ~90 m3 (6.1 m x 3.1 m x 5.5 m, Surface area to volume = 1.35 m-1) 2 mil FEP Teflon® film reactors, a 200 kW Argon arc lamp, a bank of 72 W 4-ft blacklights, along with the light and aerosol instrumentation. A schematic of the enclosure is provided in Figure 1.

Movable top frame allows reactors to collapse under pressure control

2 Banks of Blacklights This volume kept clear to maintain light uniformity

200 KW Arc Light

Dual Teflon Reactors

20 Two air Handlers ft. are located in the corners on each side of the light (not shown).

Mixing System Under floor of reactors Floor Frame 20 ft.

Temperature controlled room flushed with purified air and with reflective material on all inner surfaces

Figure 1.

20 ft.

Access Door

SEMS (PM) Instrument

Gas sample lines to laboratory below

Schematic of the environmental chamber reactors and enclosure.

The UCR EPA Environmental Chamber

29

Enclosure The interior of the thermally insulated 450 m3 enclosure is lined with reflective Everbrite® (Alcoa, PA) aluminium sheeting to maximize the interior light intensity and homogenize the interior light intensity. A positive pressure is maintained between the enclosure and the surrounding room to reduce contamination of the reactor enclosure by the surrounding building air. The enclosure air is well mixed by the large air handlers that draw in air from inlets around the light and force the air through a false ceiling with perforated reflective aluminium sheets. The enclosure is temperature controlled with a ~30 ton air conditioner capable of producing a temperature range of -5 to 80 C. Teflon Reactors The 2 mil (54 Pm) FEP Teflon® reactors are mounted within the enclosure with a rigid bottom frame and a moveable top frame. The floor of the reactor is lined with Teflon film with openings for reactant mixing within and between reactors and 8 ports ranging in size from ¼” to ½” for sample injection and withdrawal. The moveable top frame is raised/lowered with a motorized pulley system enabling the user to expand (during filling) and contract (during an experiment or for flushing) as necessary. The rate of contraction/expansion is set to maintain a differential pressure of 0.02 in H2O between the inside of the reactor and the enclosure. During experiments, the top frames are slowly lowered to maintain positive pressure as the volume decreases due to sampling, leaks, and permeation. The experiment is terminated when the final reactor volume reaches 1/3 of its maximum value (typically about 10 hours). The elevator system coupled with differential pressure measurements allows for repeatable initial chamber volumes and allows for reactants to be injected with greater than 5% precision. The Teflon reactors are built in-house using a PI-G36 Pac Impulse Sealer (San Rafael, CA) heat sealing device for all major seams and a U-frame mount to the reactor ceiling and floor. Pure Air System An Aadco 737 series (Cleves, Ohio) air purification system produces compressed air at rates up to 1500 L min-1. The air is further purified by passing through canisters of Purafil® and heated Carulite 300® followed by a filter pack to remove all particulate. The purified air within the reactor has no detectable non-methane hydrocarbons (<1 ppb), NOx (<15 ppt), no detectable particles (<0.2 particles cm-3), and a dew-point below -40 C. The reactors are cleaned between runs by reducing the reactor volume to less than 5% of its original volume while flushing the reactors with 500 L min-1 purified air. The reactor is then returned to its original volume by filling with purified air. No residual hydrocarbons, NOx, or particles are detected after the cleaning process. Light sources A 200 kW Argon arc lamp with a spectral filter (Vortek co, British Columbia, Canada) is used to irradiate the enclosure and closely simulate the entire UV-Vis ground-level solar spectra. The arc lamp is mounted on the far wall from the reactors at a minimum distance of 20’ to provide uniform lighting within both reactors. Backup lighting is provided by banks of 80 1.22 m 115-W Sylvania 350BL blacklamps (peak intensity at 350 nm) mounted on the same wall of the enclosure. These provide a low-cost and efficient UV irradiation source

W. P. L. Carter

30

Relative Intensity (Normalized to give same NO2 photolysis rate)

within the reactor for experiments where the closer spectral match provided by the Argon arc system is not required. The blacklights are located on the ends of the same wall of the Argon arc lamp. Figure 2 shows the spectra for each light source with a comparison to a representative ground-level outdoor solar spectrum.

0.12

0.8

0.10 0.6

0.08 0.06

0.4 UCR EPA Blacklights Solar Z=0

0.04 0.2

0.02 0.00 0.290

0.300

0.310

0.320

0.0 0.300

0.350

0.400

0.450

0.500

0.550

0.600

Wavelength (nm)

Figure 2 .

Spectrum of the argon arc light source used in the chamber. Blacklight and representative solar spectra, with relative intensities normalized to give the same NO2 photolysis rate.

Interreactor and Intrareactor mixing The two reactors are connected to each other through a series of custom solenoid valves and blowers. The system provides for rapid air exchange prior to the start of an experiment ensuring, when desired, that both reactors have identical concentrations of starting material. Each reactor can be premixed prior to the start of an experiment by Teflon coated fans located within the reactor. Analytical Instrumentation The instrumentation used to monitor gas-phase species in this chamber are described by Carter. (2002). Briefly, ozone, CO, NO, and NOy are monitored using commercially available instruments. Two Tuneable Diode Laser Absorption Spectroscopy (TDLAS) are available for monitoring NO2 HNO3, H2O2 and formaldehyde. TDLAS analysis is described in detail elsewhere (Hastie et al., 1983; Schiff et al., 1994) and is based on measuring single rotational - vibrational lines of the target molecules in the near to mid infrared using laser diodes with very narrow line widths and tunability. Two GC-luminol instruments based on previous work by Gaffney et al. (1998) are used to monitor PAN and higher PAN analogues. It uses luminol detection combined with GC column to separate NO2 from PAN and other species that are detected by luminol to provide a specific analysis for those species. Organic reactants other than formaldehyde are monitored by gas chromatography with FID detection.

The UCR EPA Environmental Chamber

31

Instrumentation was also available to measure aerosol size distributions, from which aerosol number and volume can be calculated. This is described by Cocker (2004), and also by Carter (2002). The use of this chamber for aerosol studies is discussed by Cocker (2004). Characterization Results Light Characterization Argon Arc Light. Although the intensity of the argon arc light can be varied by varying the lamp power, normally it is operated at 80% of maximum. Information about trends in light intensity with time is available from data from the spectral radiometer and PAR radiation instruments (Carter et al., 2002), and from results of NO2 actinometry experiments carried out periodically using the quartz tube method of Zafonte et al. (1977) modified as discussed by Carter et al. (1995a). The results indicated no significant change of light intensity with time. Actinometry experiments with the quartz tube located inside the reactors yielded an NO2 photolysis rate of 0.26 min-1. Blacklights. The light intensity inside the reactors with blacklight irradiation was measured in a series of experiments carried out in April-May of 2003 and again in October of that year, and the averages of the results were 0.19 and 0.18 min-1, respectively. The light intensity measured in the enclosure during blacklight experiments indicated a gradual decreasing trend in light intensity during the experiments that was consistent with the differences between these two measurements. Spectra. The spectrum of the arc light sources were measured using a LiCor LI-1800 spectroradiometer, and are shown on Figure 2. No appreciable change in the light source spectrum was observed in the first 18 months of operation. Characterization of Contamination by Outside Air Contamination of the reactor by leaks and permeation of laboratory air contaminants is minimized by continuously flushing the enclosure that houses the reactors with purified air. NOx and formaldehyde levels in the enclosure before or during irradiations were generally less than 5 ppb and PM concentrations are below the detection limits of our instrumentation. Introduction of contaminants into the reactor is also minimized by use of pressure control to assure that the reactors are always held at slight positive pressures with respect to the enclosure, so leaks are manifested by reduction of the reactor volume rather than dilution of the reactor by enclosure air. The leak rate into the chamber was tested by injecting ~100 ppm of CO into the enclosure and monitoring CO within the reactor. No appreciable CO (below the 50 ppb detection limit) was obtained for this experiment. Chamber Effects Characterization It is critical to characterize the impact of reactor walls on gas-phase reactivity and secondary aerosol formation. Larger volume reactors may minimize these effects but they cannot be eliminated entirely or made negligible. The most important of these for mechanism evaluation include background offgasing of NOx and other reactive species, offgasing or heterogeneous reactions that cause “chamber radical sources” upon irradiation (e.g., see Carter and Lurmann, 1990, 1991), and ozone losses to the reactor walls. Most of these can be assessed by conducting various types of characterization experiments that either directly

W. P. L. Carter

32

measure the parameter of interest, or are highly sensitive to the chamber effect being assessed. The chamber effects that have been assessed and the types of experiments that have been utilized for assessing these effects in this chamber are summarized on Table 1 and discussed further below.

Table 1.

Summary of types of characterization experiments and types of chamber effects parameters derived from these experiments.

Run Type

No. Sensitive Runs Parameters

Comments

Ozone Dark Decay

4

O3 wall loss rate

The loss of O3 in the dark is attributed entirely to a unimolecular wall loss process.

CO - Air

8

NOx offgasing

Insensitive to radical source parameters but O3 formation is very sensitive to O3 offgasing rates. Formaldehyde data can also be used to derive formaldehyde offgasing rates.

CO - HCHO air

2

NOx offgasing.

Insensitive to radical source parameters but O3 formation is very sensitive to NOx offgasing rates. Also can be used to obtain formaldehyde photolysis rates

CO - NOx

6

Initial HONO, Radical source

O3 formation and NO oxidation rates are very sensitive to radical source but not sensitive to NOx offgasing parameters. Formaldehyde data can also be used to derive formaldehyde offgasing rates.

n-Butane - NOx

1

Initial HONO, Radical source

O3 formation and NO oxidation rates are very sensitive to radical source but not sensitive to NOx offgasing parameters.

Note that some of the chamber characterization parameters are derived by conducting model simulations of the appropriate characterization experiments to determine which parameter values best fit the data. All the characterization simulations discussed here were carried out using the SAPRC-99 chemical mechanism (Carter, 2000) with the photolysis and thermal rate constants calculated using the light characterization data, the measured temperatures of the experiments, and assuming no dilution for reasons discussed above. The rates of heterogeneous reactions not discussed below, such as N2O5 hydrolysis to HNO3 or NO2 hydrolysis to HONO, were derived or estimated based on laboratory studies or other considerations as discussed previously (e.g., Carter et al., 1995a). Although the assumed values of these parameters can affect model simulations under some conditions, they are not considered to be of primary importance in affecting simulations of the characterization or other experiments discussed here.

The UCR EPA Environmental Chamber

33

NOx offgasing NOx offgasing is the main factor limiting the utility of the chamber for conducting experiments under low NOx conditions. Although this can be derived by directly measuring increases in NOx species during experiments when NOx is not injected, the most sensitive measure is the formation of O3 in irradiations when VOCs but not NOx are initially present. The NOx offgasing rate is not determine directly, but derived by determining the NOx offgasing rates in model simulations of the experiments that fit the O3 yields obtained. The NOx offgasing can be represented in the model as inputs of any species that rapidly forms NOx in atmospheric irradiation systems, such as NO, NO2, or HONO (which rapidly photolyzes to form NO, along with OH radicals), but for reasons discussed below it is represented in our chamber effects model as offgasing of HONO, e.g., Walls + hQ o HONO

Rate = k1 x RN (1)

Where k1 is the light intensity as measured by the NO2 photolysis rate, and RN is the NOx (and radical) offgasing parameter, which is derived by model simulations of the appropriate characterization experiments to determine which value best fits the data. The NOx offgasing rates that fit the results of relevant experiments carried out in the first eight months of operation of this chamber are shown as the diamond symbols on Figure 3. It can be seen that the rates were around 1.5 ppt/min up to around run 85, then increased to 2-7 ppt/min after that, being somewhat higher in the “A” compared to the “B” reactor. The reason for this increase is unclear, but it may be related to the fact that maintenance was done to the reactors around the time of the change. The magnitudes of these apparent NOx offgasing rates are discussed further below.

NOx or Radical Input (ppt/min)

7 6 Radical Input (Side A)

5

Radical input (Side B)

4

NOx Input (Side A)

NOx Input (Side B)

3

Used in Model (Side A) 2

Used in Model (Side B)

1 0 50

75

100

125

150

175

EPA Run Number

Figure 3 .

Plots of apparent NOx or radical input rates against run number.

Chamber radical source It has been known for some time that environmental chamber experiments could not be modeled consistently unless some sources of radicals attributed to chamber effects is assumed (e.g., Carter et al., 1982; Carter and Lurmann, 1991; Carter, 2000). The most sensitive

34

W. P. L. Carter

experiments to this effect are NOx -air irradiations of compounds, such as CO or alkanes, which are not radical initiators and do not form radical initiating products to a sufficient extent to significantly affect their photooxidations. In some chambers at least part of the chamberdependent radical source can be attributed to formaldehyde offgasing (Simonaitis et al., 1997, Carter, 2004a), but as discussed below the magnitude of the formaldehyde offgasing in this chamber is not sufficient by itself for the model to simulate radical-source characterization experiments. For this chamber, assuming HONO offgasing at a similar magnitude as the apparent NOx offgasing rate derived as discussed above is usually sufficient to account for most of the chamber-dependent radical source, though results of some of the experiments are somewhat better simulated if a small amount (100 ppt or less) of HONO is also assumed to be initially present. The round symbols on Figure 3 shows plots of the HONO offgasing rates that are adjusted to fit the radical-source sensitive CO - NOx and n-butane - NOx experiments that were carried out in January-October of 2003. Note that since these experiments had initial NOx levels ranging from 10 - 200 ppb, they were not sensitive to NOx offgasing as such. However, from Figure 3 it can be seen that the magnitudes of the NOx offgasing and continuous radical input rates that fit the data for the respective characterization experiments were in the same range, and even changed at the same time when the characteristics of the chamber apparently changed. Whatever effect or contamination caused the apparent NOx offgasing to increase around the time of run 85 caused the same increase in the apparent radical source. Comparison of Radical Source and NOx Offgasing with Other Chambers Although HONO is not measured directly in our experiments, the fact that both the radical-sensitive and NOx-sensitive characterization experiments can be simulated assuming HONO offgasing at approximately the same rates is highly suggestive that this is the process responsible for both effects. Direct evidence for this comes from the data of Kleffmann et al. (2003), who used sensitive long path absorption photometer (LOPAP) instrument to detect ppt levels of HONO emitted from the walls during irradiations in the large outdoor SAPHIR (Brauers et al, 2003) at rates comparable to those observed in the earlier experiments in this chamber. The SAPHIR chamber is similar in design to the chamber discussed here, except it is larger in volume and is located outdoors. In particular, like our chamber it has Teflon walls and uses a dual-reactor configuration to minimize contamination by outside air. Therefore, it would be expected to have similar chamber NOx and radical sources, and this appears to be the case. Figure 4 shows plots of the NOx offgasing or radical source parameter (e.g. RN in Equation 1) obtained in modeling appropriate characterization runs in various chambers, where they are compared with direct measurements made in the SAPHIR chamber (Kleffmann et al. (2003). In addition to this UCR EPA chamber, the radical source parameters shown are those derived by Carter (2000) for previous indoor and outdoor chambers at UCR (Carter et al., 1995a), those derived by Carter and Lurmann (1990, 1991) for the University of North Carolina (UNC) outdoor chamber (e,g., Jeffries et al., 1982, 1990), and those derived by Carter (2004a) for the Tennessee Valley Authority (TVA) indoor chamber (Simonaitis and Bailey, 1995; Bailey et al., 1996). (Note that the data shown for the UCR EPA chamber includes experiments carried out subsequently to those shown on Figure 3, including a few runs at reduced temperature.) The figure shows that the radical source and NOx offgasing rates derived for this chamber are comparable in magnitude to the HONO offgasing directly

The UCR EPA Environmental Chamber

35

measured in the SAPHIR chamber and also comparable to the NOx offgasing derived for TVA chamber but are significantly lower than those derived from modeling characterization data from the earlier UCR and UNC chambers. It is interesting to note that parameters derived for the various chambers all indicate that the radical source and HONO or NOx offgasing rates all increase with temperature.

1000 NOx offgasing or Radical Source / NO2 photolysis rate (ppt)

Previous UCR Indoor Previous UCR Outdoor

Previous UCR Default Model 100

UNC Outdoor Model

TVA Indoor (NOx offgasing) UCR EPA (first series)

10

UCR EPA (later series) SAPHIR HONO offgasing (dry) 1 290

300

310

320

Average Temperature (K)

Figure 4 .

Plots of ratios of NOx offgasing or radical source to the NO2 photolysis rates derived from modelling characterization runs for various chambers.

Therefore, the radical source and NOx offgasing rates indicated by the characterization data for the first series of experiments for this chamber is probably as low as one can obtain for reactors constructed of FEP Teflon film, which is generally believed to be the most inert material that is practical for use as chamber walls. Although the radical source and NOx offgasing rates for the second series of experiments is higher (see also Figure 3), they are still about an order of magnitude lower than observed for the UCR and UNC chambers previously used for mechanism evaluation. Formaldehyde offgasing Low but measurable amounts of formaldehyde were formed in irradiations in this chamber, even in pure air, CO - NOx, or other experiments where no formaldehyde or formaldehyde precursors were injected. Some formaldehyde formation is expected from the reactions of methane in the matrix air, since the methane removal catalyst was not operational during the period of these experiments, but formaldehyde formation in this reaction in the CO experiments is predicted to be negligible. The data in essentially all such experiments could be modelled assuming a continuous light-dependent formaldehyde offgasing rate corresponding to 7.5 ppb/day at the light intensity of these experiments. This is a relatively low offgasing rate that could not be detected with formaldehyde analyzers used in most previous UCR and other chamber experiments, and is insufficient to account for the apparent chamber radical

36

W. P. L. Carter

source observed in most chamber experiments. This apparent formaldehyde offgasing has a non-negligible effect on very low VOC and radical source characterization experiments, so it must be included in the characterization model. However, it has a relatively minor impact on modelling most experiments used for VOC mechanism evaluation or reactivity assessment. The source of the apparent formaldehyde offgasing in the Teflon reactors is unknown, but it is unlikely to be due to build-up of contaminants from previous exposures or contamination from the enclosure. The apparent formaldehyde offgasing rate is quite consistent in most cases and there are no measurable differences between the two reactors or changes in formaldehyde offgasing with time. The data are best modelled by assuming only direct formaldehyde offgasing, as opposed to some formaldehyde precursor being formed from light-induced reactions of some undetected contaminant, as is apparently important in the TVA chamber (Simonaitis et al., 1997; Carter, 2004a). Representative Results of Initial Evaluation Experiments Table 2 gives a summary of the initial experiments carried out in this chamber for chamber and mechanism evaluation. The characterization and single organic - NOx experiments and some of the ambient reactive organic gas (ROG) surrogate - NOx experiments were funded by the EPA cooperative agreement used to fund the construction and initial characterization of this chamber. The lowest concentration ambient ROG surrogate NOx experiments were carried out by a separate California Air Resources Board project to obtain additional data for low NOx mechanism evaluation, and a number of surrogate - NOx experiments at varying initial ROG and NOx levels at a were carried out for a separate program to obtain data to evaluate observationally based methods. Not shown on Table 2 are a number of “incremental reactivity” experiments, where various compounds or VOC mixtures were added to standard ambient ROG - NOx surrogate irradiations to provide data to test mechanisms for their relative ozone impacts. The results of these experiments are still being analyzed. The various characterization experiments were used to derive the chamber characterization parameters and evaluate the chamber characterization model as discussed above. The single organic - NOx experiments were carried out to demonstrate the utility of the chamber to test the mechanisms for these compounds, for which data are available in other chambers, and to obtain well-characterized mechanism evaluation data at lower NOx levels than previously available. The formaldehyde + CO - NOx experiments were carried out because they provided the most chemically simple system that model calculations indicated was insensitive to chamber effects, to provide a test for both the basic mechanism and the light characterization assignments. The aromatic - CO - NOx experiments were carried out because aromatic - NOx experiments were predicted to be very sensitive to the addition of CO, because it enhances the effects of radicals formed in the aromatic system on ozone formation. The ambient surrogate - NOx experiments were carried out to test the ability of the mechanism to simulate ozone formation under simulated ambient conditions at various ROG and NOx levels.

The UCR EPA Environmental Chamber

Table 2.

37

Summary of initial chamber and mechanism evaluation experiments.

Runs

NOx Range (ppb)

VOC Range (ppm)

Avg. Model Bias #

Avg. Model Error #

Characterization

32

0 - 200

Varied

-3%

28%

Formaldehyde - NOx

2

8 - 25

0.35 – 0.50

-23%

23%

Formaldehyde - CO - NOx

2

15 - 20

HCHO: 0.4 – 0.5 CO: 15 - 80

-10%

10%

Ethene - NOx

2

10 - 25

~0.6

-15%

15%

Propene - NOx

2

5 - 25

0.4 – 0.5

16%

16%

Toluene or m-Xylene - NOx

4

5 - 25

10%

10%

Aromatic - NOx + CO

6

5 - 30

Toluene: 0.6 – 0.16 m-Xylene: 0.18 CO: 25 – 50

-17%

18%

61*

2 - 315

0.2 – 4.2 ppmC

-10%

13%

Run Type

Ambient Surrogate - NOx

* Includes experiments carried out for subsequent programs. # Error and bias for model predictions of '([O3]-[NO]) using the SAPRC-99 mechanism. Bias is (calculated - experimental) / calculated. Error is the absolute value of the bias.

The ROG surrogate used in the ambient surrogate - NOx experiments consisted of a simplified mixture of designed to represent the major classes of hydrocarbons and aldehydes measured in ambient urban atmospheres, with one compound used to represent each model species used in condensed lumped-molecule mechanism. The eight representative compounds used were n-butane, n-octane, ethene, propene, trans-2-butene, toluene, m-xylene, and formaldehyde. (See Carter et al., 1995b, for a discussion of the derivation of this surrogate). The ability of the SAPRC-99 mechanism to simulate the total amount of NO oxidized and O3 formed in the experiments, measured by ([O3]final-[NO]final) - ([O3]initial-[NO]initial) or '([O3]-[NO]), is summarized for the various types of experiments on Table 2 and shown for the individual runs on Figure 5. This gives an indication of the biases and run-to-run variability of the mechanism in simulating ozone formation. Note that in experiments with excess NO the processes responsible for O3 formation are manifested by consumption of NO, so simulations of '([O3]-[NO]) provides a test of model simulations of these processes even for experiments where O3 is not formed. Note that the characterization runs were modelled using the default characterization parameters used when modelling the mechanism evaluation runs, not with the values that were adjusted to fit the individual experiments. Therefore, the relatively large variability and average model error for the model simulations of '([O3]-[NO]) in those experiments provides a measure of the variability of the chamber effects parameters (e.g., HONO offgasing) to which these experiments are sensitive. The relatively low average bias is expected because the chamber effects parameter values were derived based on these data.

W. P. L. Carter

38 Single Run

Average

All Characterization HCHO - NOx HCHO - CO - NOx Ethene - NOx Propene - NOx Toluene - NOx Toluene - CO - NOx m-Xylene - NOx m-Xylene - CO - NOx Surrogate - NOx -75%

-50%

-25%

0%

25%

50%

75%

(Calculated - Experimental) / Experimental

Figure 5 .

Fits of experimental '([O3]-[NO]) measurements to SAPRC-99 model calculations for the initial chamber and mechanism evaluation experiments.

For the single VOC - NOx or VOC - CO - NOx experiments, the model is able to simulate the '([O3]-[NO]) to within r25% or better in most cases, which is better than the r~30% seen in previous mechanism evaluations with the older chamber data (Carter and Lurmann, 1990, 1991; Gery et al., 1989, Carter, 2000). However, there are indications of nonnegligible biases in model simulations of certain classes of experiments. The cleaner conditions and the relatively lower magnitude of the chamber effects may make the run-to-run scatter in the model performance may be less than in the simulations of the previous data, and this tends to make relatively small biases in the model performance more evident. There are, for example, definite biases in the model to under predict O3 formation and NO oxidation in the surrogate - NOx experiments carried out at lower ROG/NOx ratios. These cases are discussed further elsewhere (Carter, 2004b). As indicated on Table 2, the initial evaluation experiments included runs with NOx levels as low as 2-5 ppb, which is considerably lower than in experiments used previously for mechanism evaluation. Most of the experiments used in the previous SAPRC-99 mechanism evaluation had NOx levels greater than 50 ppb, and even the “low NOx” TVA and CSIRO experiments had NOx levels of ~20 ppb or greater, except for a few characterization runs (Carter, 2004a, and references therein). However, there is no indication in any difference in model performance in simulating the results of these very low NOx experiments, compared to those with the higher NOx levels more representative of those used in the previous evaluation. For example, Figure 6 shows concentration-time plots for selected measured species in ambient surrogate - NOx experiment carried out at the lowest NOx levels in the initial evaluation runs. To indicate the sensitivity of the experiments to NOx offgasing effects, the effects of varying the HONO offgasing parameter from zero to the maximum level consistent with the characterization experiments is also shown. It can be seen that the model using the default HONO offgasing parameter value gives very good fits to the data. Although the O3 simulations are somewhat affected when the HONO offgasing rate is varied within this somewhat extreme range, the sensitivity is not so great that the uncertainty in this parameter

The UCR EPA Environmental Chamber

39

significantly affects conclusions one can draw about the ability of the model to simulate this low NOx experiment. However, the sensitivity would increase as the NOx levels are reduced, and ~2 ppb NOx probably represents a reasonable lower limit for NOx levels useful for mechanism evaluation. O3

0.04

NO

0.002

NO2

0.0015

0.03

0.0010

0.001

0.02

0.0005

0.01 0.00

0.000

0

120

240

360

m-Xylene

0.0000

0

120

240

360

120

240

360

240

360

PAN

NOy - HNO3

0.004

0

0.0006

0.006

0.003 0.004

0.0004

0.002

0.002

0.0002

0.001 0.000

0.000 0

120

240

360

0.0000

0

120

240

360

0

120

Concentration (ppm) vs Time (minutes)

Experimental No HONO Offgasing

Figure 6.

Standard Model Maximum HONO Offgasing

Concentration-time plots of selected compounds in the lowest NOx ambient ROG - NOx surrogate experiment in the initial evaluation experiments (NOx | 1 ppb, ROG | 300 ppbC.

Overall, the results of the initial characterization and evaluation indicate that this chamber can provide high quality mechanism evaluation data for experiments with NOx levels as low as ~2 ppb, considerably lower than employed in previous experiments. Chamber effects are not absent, but they are as low as or lower than in observed in any previous chambers used for mechanism evaluation, in some cases by an order of magnitude or more. Although a larger number of experiments would be required to fully assess this, the results also suggest a higher degree of precision in mechanism evaluation than observed previously, making smaller biases in mechanism performance more evident. The initial dataset from this chamber indicate no significant problems with mechanism performance that are characteristic of low NOx conditions as such, but do reveal problems with the mechanisms for aromatics and the ambient ROG surrogate. This is discussed further in our companion presentation (Carter, 2004b). Future Research Directions The future research directions for the UCR EPA chamber will of course depend on funding availability and the interests of the funding agencies, but we would expect the research would be designed to take full advantage of its capabilities. In any case, we expect to continue the ozone reactivity and gas-phase mechanism studies that are currently underway, with the mechanism evaluation being focused on aromatics, whose mechanism are the most

40

W. P. L. Carter

uncertain and where the new data indicate mechanism problems (Carter, 2004a,b), and ozone reactivities of VOC source categories of interest. We will also be utilizing the chambers for initial experiments to investigate temperature effects on O3 formation and other measures of gas-phase reactivity, taking advantage of the temperature control capabilities of this chamber, which has not yet been fully exploited. We hope in the future to obtain instrumentation needed for NO3, N2O5, HOx, and other trace species in order to improve capabilities and utility of this facility for mechanism evaluation. As discussed in more detail by Cocker (2004), we will also utilize the chamber for well-characterized experiments needed to develop and evaluate models for secondary organic aerosol (SOA) formation. The temperature control capabilities of this chamber permit systematic studies of temperature and humidity effects that are not possible with existing outdoor chambers or with indoor chambers using blacklights (whose intensities are temperature-dependent). The established capability of this chamber for gas phase mechanism evaluation is an important characteristic in this regard, since the model cannot properly simulate SOA formation if it does not also properly simulate the formation of the SOA precursors from the gas-phase reactions. To be fully utilized, this chamber should also serve as a resource for collaborative studies where environmental chamber measurements under highly controlled conditions would be useful. We encourage collaboration with other researchers who have instrumentation that we lack that would be useful for various studies, such as our collaboration with William Brune of Penn State University, whose HO and HO2 radical measurement instrumentation was used with the surrogate - NOx experiments carried out for the observational based methods study. This facility would also be useful as a test bed for evaluating analytical instrumentation for use in ambient monitoring, since it can produce realistic but highly characterized simulated pollution conditions for instrument evaluation and inter-comparisons. References Bailey, E. M., C. H. Copeland and R. Simonaitis; Smog chamber studies at low VOC and NOx concentrations, Report on Interagency Agreement DW64936024 to EPA/NREL, Research Triangle Park, NC. (1996). Brauers, T., B. Bohn, F.-J. Johnen, F. Rohrer, S. Rodriguez Bares, R. Tillmann, and A. Wahner; The atmosphere simulation chamber SAPHIR: a tool for the investigation of photochemistry, Presented at the EGS AGU - EUG Joint Assembly, Nice, France, April 11 (2003). See also http://www.fz-juelich.de/icg/icgii/saphir/home. Carter, W. P. L.; Documentation of the SAPRC-99 chemical mechanism for VOC reactivity assessment, Report to the California Air Resources Board, Contracts 92-329 and 95-308, May 8 (2000). Available at http://www.cert.ucr.edu/~carter/absts.htm#saprc99. Carter, W. P. L.; Development of a next-generation environmental chamber facility for chemical mechanism and VOC reactivity research. Draft research plan and first progress report to the United States Environmental Protection Agency Cooperative Agreement CR 827331-01-0. January 3 (2002). Available at http://www.cert.ucr.edu/~carter/epacham. Carter, W. P. L.; Evaluation of a gas-phase atmospheric reaction mechanism for low NOx conditions, Final Report to California Air Resources Board Contract No. 01-305, May 5 (2004a). Available at http://www.cert.ucr.edu/~carter/absts.htm#lnoxrptf Carter, W. P. L.; Environmental chamber studies of ozone formation potentials of VOCs. Presented at the NATO EST-ARW Workshop on Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, Zakopane, Poland, October 1-4 (2004b).

The UCR EPA Environmental Chamber

41

Carter, W. P. L., R. Atkinson, A. M. Winer, and J. N. Pitts, Jr.; Experimental Investigation of ChamberDependent Radical Sources, Int. J. Chem. Kinet., 14 (1982) 1071. Carter, W. P. L., and F. W. Lurmann; Evaluation of the RADM gas-phase chemical mechanism. Final Report to the U.S. Environmental Protection Agency, EPA-600/3-90-001 (1990). Carter, W. P .L., and F. W. Lurmann; Evaluation of a detailed gas-phase atmospheric reaction mechanism using environmental chamber data. Atmos. Environ. 25A (1991) 2771-2806. Carter, W. P. L., D. Luo, I. L. Malkina, and D. Fitz; The University of California, Riverside environmental chamber data base for evaluating oxidant mechanism. Indoor chamber experiments through 1993, Report submitted to the U. S. Environmental Protection Agency, EPA/AREAL, Research Triangle Park, NC., March 20 (1995a). Available at http://www.cert.ucr.edu/~carter/ absts.htm#databas. Carter, W. P. L., D. Luo, I. L. Malkina, and J. A. Pierce; Environmental chamber studies of atmospheric reactivities of volatile organic compounds. Effects of varying ROG surrogate and NOx, Final report to Coordinating Research Council, Inc., Project ME-9, California Air Resources Board, Contract A0320692, and South Coast Air Quality Management District, Contract C91323. March 24 (1995b). Available at http://www.cert.ucr.edu/~carter/absts.htm#rct2rept. Cocker, D. R. State-of-the art chamber for aerosol studies. Presented at the NATO EST-ARW Workshop on Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, Zakopane, Poland, October 1-4 (2004). Dodge, M. C.; Chemical oxidant mechanisms for air quality modeling, critical review paper for 1998 Ozone Assessment, Atmos. Environ, 34 (2000) 2103-2130. Gaffney, J. S, R. M. Bornick., Y-H Chen, and N. A. Marley; Capillary gas chromatographic analysis of nitrogen dioxide and PANs with luminol chemiluminescent detection, Atmos. Environ. 32 (1998) 1445-1454. Gery, M. W., G. Z. Whitten, J. P. Killus, and M. C. Dodge; A Photochemical Mechanism for Urban and Regional Scale Computer Modeling, , J. Geophys. Res, 94, (1989) 12,925. Hastie, D. R, G. I. Mackay, T. Iguchi, B. A. Ridley, and H. I. Schiff, H. I; Tunable diode laser systems for measuring trace gases in tropospheric air, Environ. Sci. Technol. 17, (1983) 352A-364A. Jeffries, H. E., R. M. Kamens, K. G. Sexton, and A. A. Gerhardt; outdoor smog chamber experiments to test photochemical models, EPA-600/3-82-016a, April (1982). Jeffries, H. E., K. G. Sexton, J. R. Arnold, Y. Bai, J. L. Li, and R. Crouse; A Chamber and Modeling Study to Assess the Photochemistry of Formaldehyde, Report on EPA Cooperative Agreement CR-813964, Atmospheric Research and Exposure Assessment Laboratory, EPA, Research Triangle Park, NC. (1990). Jeffries, H.E.; M. W. Gery, and W. P. L. Carter; Protocol for evaluating oxidant mechanisms for urban and regional models. Report for U.S. Environmental Protection Agency Cooperative Agreement No. 815779, Atmospheric Research and Exposure Assessment Laboratory, Research Triangle Park, NC. (1992). Kleffmann, J., B. Bohn, T. Brauers, F. Rohrer, A. Wahner, and P. Wiesen: Photolytic HONO formation in the atmospheric simulation chamber SAPHIR. Presented at the EGS - AGU - EUG Joint Assembly, Nice, France, April 11 (2003). Schiff, H.. I., G. I. Mackay, and J. Bechara; The use of tunable diode laser absorption spectroscopy for atmospheric measurements, Res. Chem. Intermed. 20, 1994 (1994) 525-556. Simonaitis, R. and E. M. Bailey; Smog chamber studies at low VOC and NOx concentrations: Phase I, Report on Interagency Agreement DW64936024 to EPA/NREL, Research Triangle Park, NC. (1995). Simonaitis, R., J. Meagher, and E. M. Bailey; Evaluation of the condensed Carbon Bond Mechanism against smog chamber data at low VOC and NOx Concentrations, Atm. Environ. 31 (1997) 27-43. Zafonte, L., P. L. Rieger, and J. R. Holmes; Nitrogen dioxide photolysis in the Los Angeles atmosphere, Environ. Sci. Technol. 11 (1977) 483-487.

Investigations of Secondary Organic Aerosol in the UCR EPA Environmental Chamber David R. Cocker III and Chen Song Department of Chemical and Environmental Engineering, UC Riverside, Riverside, California, 92521, USA Key Words: Environmental chambers, Secondary organic aerosol, Oxides of nitrogen, m-Xylene

Abstract A large indoor environmental reactor was completed in 2002 at UC Riverside to investigate photochemical processes leading to ozone and secondary organic aerosol (SOA) formation. The chamber is briefly described by Carter (2004) in a separate contribution to this workshop and in a more detailed manuscript submitted to Atmospheric Environment (Carter et al., 2005). The contribution to the workshop covered some details of the UCR chamber wall characterization (Carter, 2004), selected approaches to studies of secondary organic aerosol formation (Cocker et al., 2001 and Song et al., 2005), the NOx dependence of m-xylene SOA formation (Song et al, 2005), some acid catalyzed reaction work from Caltech (Gao et al., 2004), and select details of the Caltech environmental chamber (Cocker et al., 2001). This contribution will briefly describe the SOA components of the chamber facility and advantages of the new reactor with respect to SOA investigations, and will summarize the findings on the importance of hydrocarbon:NOx ratio on SOA formation in the m-xylene photooxidation system. Introduction The environmental chamber can be considered to be a well-mixed batch reactor where one can study atmospheric processes in the absence of microphysical transport processes. Volumes for such reactors range from tens of liters to hundreds of cubic meters with a variety of surfaces selected to minimize the impact of the walls on the chemistry occurring within the reactor (Becker, 2004 and references therein). Secondary organic aerosol remains a significant contributor to regional air pollution with estimates of up to 80% of fine particulate being secondary in origin in heavily impacted regions (Turpin et al., 1991 and Na et al., 2004) and may play an important role in visibility degradation, global climate change, and adversely affect human health (Seinfeld, 2004 and references therein). Traditionally, the extent of secondary organic aerosol formation from a reacted hydrocarbon precursor has been discussed in terms of aerosol yield (Y), defined as the mass concentration of aerosol formed divided by the mass concentration of reacted hydrocarbons. The aerosol yield is then described using a two-parameter semi-empirical approach (Pankow, 1994a,b and Odum et al., 1996 ) Y

¦Y

i

i

'M org ¦ i

D i Ȁ om,i

(1)

1  Ȁ om,i 'M org 43

I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 43–47. © 2006 Springer. Printed in the Netherlands.

D. R. Cocker III and C. Song

44

where Yi is the yield of individual aerosol forming products, Di is the mass-based stoichiometric fraction of species i formed from the parent hydrocarbon, Kom,i is the gasparticle partitioning coefficient (m3Pg-1), which is inversely proportional to the compound’s vapor pressure, and 'Morg (Pgm-3) is the total mass concentration of organic material and associated water present in the aerosol phase. The fraction of secondary organic material condensing into the aerosol phase is seen to depend on the amount of organic aerosol mass present. Furthermore, Kom,i can be estimated (Pankow, 1994a,b) from K om ,i

760 RTf om ,i 10 MWom] i p Lo ,i 6

(2)

where R is the ideal gas constant; T is temperature; fom is the weight fraction of the total suspended particulate that comprises the absorbing phase; MWom is the number-averaged molecular weight of the absorbing organic matter phase, ]i is the activity coefficient of species i and poL,i is the sub-cooled liquid vapor pressure. The two-product semi-empirical model then assumes that two surrogate species can be used to estimate the SOA yield: one surrogate product representing low vapor pressure compounds and one surrogate product representing high vapor pressure compounds (i=1,2 in Equation 1). Current application of the two-product model does not provide an obvious means to incorporate changes in gas-phase chemistry due to the existence of nitrogenous species, light hydrocarbons, or water. Little work has been performed to date on the effect of NOx on SOA yield. Izumi et al. (1990) performed two experiments with the same initial m-xylene concentrations but different initial NOx concentrations and found similar SOA yields. Odum et al. (1996) investigated SOA formation from m-xylene photooxidation and concluded that the SOA yield for aromatic hydrocarbons is weakly dependent on HC:NOx. Hurley et al. (2001) investigated SOA formation from toluene and concluded that SOA yield for toluene with and without ozone (high NOx) was 0.108±0.004 and 0.075±0.004, respectively. Most recently, Wirtz (2004) at this conference and Martins-Reviejo et al. (2005) reported increased aerosol formation per reacted hydrocarbon at lower NOx concentrations for benzene and toluene. This paper (published in detail by Song et al. (2005) presents significantly increased aerosol formation per reacted m-xylene as hydrocarbon to NOx ratio increases. Also presented are some details on current aerosol measurement capabilities and unique advantages for SOA work in this facility. Brief chamber description The UC Riverside facility consists of dual 90 m3 chambers attached to a movable frame that is slowly lowered to maintain a slightly positive differential pressure. The chambers are located within a large, temperature controlled refrigeration container that is continually flushed with purified dry air. Two light sources are available: a 200 kW argon arc lamp and two banks of 40 each 72 W 1.22 m fluorescent blacklamps. The temperature and humidity within the chamber can be accurately controlled (See Carter et al., 2005 for more detail). Advantages of the indoor reactor include temperature (recall from equation 2 the

Investigations of Secondary Organic Aerosol in the UCR EPA Environmental Chamber

45

importance of temperature on Kom,i) and humidity control. Additionally, the indoor reactors benefit from reproducible lighting conditions throughout an experiment and experiment to experiment and provide for improved comparison between different experiments. However, the key disadvantage is the lack of a true “solar” radiation spectrum. Spectral differences are largely overcome in this reactor due to a 200 kW argon arc lamp irradiation source that performs a reasonable simulation of the UV-visible outdoor solar spectra (see Carter, 2004). Regardless of the location of the chamber (indoor vs. outdoor), the absorption properties of the film will lead to some modification of incoming radiation. Instrumentation: Basic aerosol instrumentation is currently limited to a suite of Scanning Electrical Mobility Spectrometers (SEMS) (Wang and Flagan, 1989) that monitor particles in the 28-730 nm range. The instrument design and operation is very similar to those described by Cocker et al. (2001). Sheath air for particle classification is drawn and filtered from the atmospheric reactor to ensure that the sheath gas surrounding the particles would be similar in chemical composition thus reducing bias due to evaporation or condensation of semi-volatile material. The classification portion of the SEMS are housed within the temperature controlled enclosure to ensure that no bias occurs in sampling due to changes in sampling line temperatures. Additionally, a fast scanning mobility particle spectrometer (f-SMPS) (size range 3-100 nm, 1 second time resolution, see also Shah et al., 2005) and a tandem differential mobility analyzer (TDMA, see also Cocker et al., 2001) are available to track ultrafine particle production and aerosol hygroscopic properties, respectively. Offline instrumentation for detailed chemical characterization of gaseous, semi-volatile, and particulate products are also available. Particle loss rate: The particle loss rate for the chamber can be described as a first-order wall loss process (Cocker et al., 2001). Particle loss rates determined for the first set of aerosol experiments within the reactor is about 7 day-1. While maximum particle volume in experiments performed have ranged from less than 0.1 to almost 80 µg/m3, no correlation between maximum particle volumes and measured decay rates has been seen (Carter et al., 2005). Impact of NOx on aerosol formation A series of experiments have been performed in the UCR chamber to ascertain impacts of NOx on aerosol formation in the m-xylene/NOx experiments. A full research paper on this subject has already been published in ES&T (Song et al.) A brief summary of the work (presented in Zakopane) follows. A series of 24 m-xylene photoxidation experiments (blacklamp irradiation) were performed at 300K. Reacted hydrocarbon mass concentrations ranged from 149 to 1660 Pg m-3, NOx concentrations ranged from 9 to 587 ppb, and HC:NOx ratios (ppbC/ppb NOx) ranged from 1.3 to nearly 50. All experiments were performed until wall loss corrected aerosol volume plateaued reached a constant stable value. While the data for all experiments could be plotted as yield versus organic mass produced and generally explained by a two-product fit (0.040, 0.164, 0.250 and 0.012 for Į1,

D. R. Cocker III and C. Song

46

Į2, K1 and K2), closer inspection of the aerosol mass concentration produced for experiments with similarly reacted hydrocarbons showed significantly different total aerosol concentration. The discrepancy between quality of fit for Y vs. Mo and the realization that aerosol mass concentration produced varied significantly for similar hydrocarbon reacted was simply attributed to the fact that the y-axis was not truly independent of the x-axis. (Y is inversely proportional to Mo; Mo is the x-axis variable). Therefore, to see the effect of NOx on SOA formation more clearly, the data was replotted as Mo versus HC reacted and the two-product model solved as Mo

1 ª 2 2 K1K2'HCD1  D2  K1  K2  K12 K22'HC2 D1  D2  2K1K2'HCK1  K2 D1  D2  K1  K2 º »¼ 2K1K2 «¬

(3)

This allowed for the determination of two sets of distinct aerosol curves, one for low HC:NOx ratios (<5.5) and one for high HC:NOx ratios (>8.0). Values of Į1, Į2, K1 and K2 are 0.024, 0.152, 0.229, 0.004 for low HCo:NOx ratio experiments (R2>0.99) and 0.049, 0.178, 0.301, 0.008 for high HCo:NOx ratios (R2>0.99). (Song et al., 2005). A plot of the modeled aerosol mass concentrations versus hydrocarbon reacted can be found in Figure 1.

400 HCo:NOx<5.5 HCo:NOx>8.0

350

Mo(Pg/m3)

300 250 200 150 100 50 0 0

500

1000

1500

2000

3

HC(Pg/m ) Figure 1.

SOA mass concentrations from m-xylene photooxidation as a function of mxylene consumption. Empirical fit using equation 3 for low NOx (HCo:NOx>8.0) and high NOx (HCo:NOx<5.5) conditions.

Of critical interest is the large deviation in aerosol produced between the low NOx (conditions more typical of the ambient) and high NOx (conditions often employed in atmospheric reactors) conditions, particularly for aerosol mass concentrations and reacted hydrocarbon concentrations typical of the polluted urban atmosphere (lower left portion of the curve).

Investigations of Secondary Organic Aerosol in the UCR EPA Environmental Chamber

47

Finally, it must be noted that the discrepancy between the aerosol productions could not be attributed solely to total nitrate radical or ozone concentrations. Furthermore, the onset of aerosol formation (aerosol threshold) in the systems was found to be strongly dependent on the HC:NOx ratios present in the chamber. References Becker, K.H. Historical Overview on the Development of Chambers for the Study of Atmospheric Processes; Presented at the NATO EST-ARW Workshop on Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, Zakopane, Poland, October 1-4 (2004). Carter, W.P.L. The CE-CERT Chamber in Riverside, California; Presented at the NATO EST-ARW Workshop on Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, Zakopane, Poland, October 1-4 (2004). Carter, W.P.L., D.R. Cocker III, D.R. Fitz, I.L. Malkina, K. Bumiller, C.G. Sauer, J.T. Pisano, C. Bufalino, C. Song; A New Environmental Chamber for Evaluation of Gas-Phase Chemical Mechanisms and Secondary Aerosol Formation. Submitted to Atmospheric Environment, 2005. Cocker D.R. ɒ, R.C. Flagan, J.H. Seinfeld; State-of-the-art chamber facility for studying atmospheric aerosol chemistry. Envir. Sci. Technol. 35 (2001) 2594-2601. Gao S., N.L. Ng, M. Keywood, V. Varutbangkul, R. Bahreini, A. Nenes, J.W. He, K.Y. Yoo, J.L. Beauchamp. R.P. Hodyss, R.C. Flagan, J.H. Seinfeld; Particle phase acidity and oligomer formation in secondary organic aerosol. Envir. Sci. Technol. 38 (2004) 6582-6589. Hurley M. D., O. Sokolov, T.J. Wallington, H. Takekawa, M. Karasawa, B. Klotz, I. Barnes, and K. Becker; Organic aerosol formation during the atmospheric degradation of toluene. Environ. Sci. Technol. 35 (2001) 1358-1366. Izumi K. and T. Fukuyama; Photochemical aerosol formation from aromatic-hydrocarbons in the presence of NOx. Atmos. Environ. 24A:6 (1990), 1433-1441. Martin-Reviejo, M. and K. Wirtz. Is Benzene a Precursor for Secondary Organic Aerosol? Environ. Sci. Technol.39 (2005), 1045-1054. Na K., A.A. Sawant, C. Song, and D.R. Cocker; Primary and secondary carbonaceous species in the atmosphere of Western Riverside County, California. Atmos. Environ. 38 (2004) 1345-1355. Odum J. R., Hoffmann T., Bowman F., Collins D., Flagan R. C. and Seinfeld J. H. Gas/particle partitioning and secondary organic aerosol yields. Envir. Sci. Technol. 1996, 30, 2580-85. Pankow J. F. An absorption-model of gas-particle partitioning of organic-compounds in the atmosphere. Atmos. Environ. 1994a, Vol. 28, No. 2, 185-188. Pankow J. F. An absorption model of the gas-aerosol partitioning involved in the formation of secondary organic aerosol. Atmos. Environ.1994b, Vol. 28, No. 2, 189-193. Seinfeld, J.H, and Pankow, J.F. Organic Atmospheric Particulate Material. Annu. Rev. Phys. Chem.¸2003, Vol. 54:121-140. Shah, S.D., and Cocker III,D.R., A Fast Scanning Mobility Particle Spectrometer for Monitoring Transient Particle Size Distributions, submitted to Aero. Sci. & Technol., 2005. Song C., Na K., Cocker D. R. Impact of Hydrocarbon to NOx ratio on Secondary Organic Aerosol Formation. Envir. Sci. Technol. 2005, In Press. Turpin B. J. and Huntzicker J. J. Secondary formation of organic aerosol in the Los-angles basin – a descriptive analysis of organic and elemental carbon concentrations. Atmos. Environ. 1991, Vol. 25A, No. 2, 207215. Wang S. C., Flagan R. C. Scanning Electrical Mobility Spectrometer. Aero. Sci. Technol. 1990, 13, 230-240. Wirtz, K. Recent Investigations on Aerosol Formation in EUPHORE Chamber. Presented at the NATO ESTARW Workshop on Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, Zakopane, Poland, October 1-4 (2004).

Field Measurement and Modelling Motivated Design of a Programme of Aerosol Chamber Experiments Gordon McFiggans Atmospheric Sciences Group, School of Earth, Atmospheric & Environmental Sciences (SEAES), University of Manchester, Sackville Street Building, Manchester, M60 1QD, UK Key Words: Atmospheric aerosol, Chamber studies, Coupled modelling

Abstract This paper describes the development of a chamber facility designed to investigate a range of atmospheric aerosol processes driven by their potential to affect radiative forcing. Development of a suite of models investigating coupled chemical and microphysical processes has informed the definition of a preliminary experimental programme. The chamber construction is underway and the experiments are phased to follow the chamber development and completion. Introduction Atmospheric aerosols are the focus of vigorous and extensive current investigation due to their roles in climate change through direct and indirect radiative forcing and to their impacts on human health. This paper focuses on the development of a chamber facility to investigate aerosols and aerosol processes which may impact on both, but will be characterised in terms of their ability to participate in climate change. The climatic implications depend critically on the number of aerosols of optically active size, their composition and resulting affinity for water in the sub-saturated environment and their ability to activate into cloud droplets and ice crystals moving into supersaturation. Using climatic implications as primary development drivers it was decided to design an aerosol chamber and work programme of targeted experiments to investigate example aerosol systems and to demonstrate the suitability of such a chamber to probe the changes in radiative forcing potential with aerosol processing. In contrast to atmospheric gas phase processes, the understanding of the cycling and implications of atmospheric aerosols is relatively young and underdeveloped. This is, in part, due to the complexity of the coupling between the microphysics and chemistry and also due to the lack of availability of suitable instrumentation to probe aerosol chemical composition and physical behaviour in the field with sufficient time resolution (and even in the laboratory with sufficient selectivity). Given the current state-of-the-science, it is inevitable that field observations of aerosol physical and chemical properties regularly throw up surprises and unanswered questions. It is unclear whether the most important questions are being addressed in current detailed laboratory studies where many of the approaches have been borrowed from the gas phase laboratory work. Recent adaptation and application of techniques from surface chemistry show promise in terms of detailed mechanistic interpretation of multiphase processes, but the interpretation of findings from such studies and their incorporation into even simple atmospheric models is difficult (due to problems extrapolating from aerosol mimics, for example). In addition, the degree of complexity in gaseous composition has been widely reported (e.g. Lewis et al., 2000) and, though less well characterised, it is likely that the composition of aerosols will exhibit comparable complexity. Indeed, due to the range of 49 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 49–65. © 2006 Springer. Printed in the Netherlands.

50

G. McFiggans

sizes and possible mixing states, a detailed understanding of the intricate chemical and microphysical couplings in many aerosol systems is unlikely to be tractable, and may not even be desirable. Whilst modern kinetic flow studies are needed to understand underlying mechanistic concepts and levitation techniques can clarify single particle thermodynamic behaviour, it is difficult to directly and quantitatively evaluate the importance of such complex systems by testing models using such data directly against field measurements. Chamber facilities have played a central role in recent developments in gas phase atmospheric chemistry, allowing the development of understanding of complex mechanisms and they are increasingly used to investigate the formation and evolution of aerosols. The main advantage of chamber studies is the ability to separate chemical and microphysical atmospheric processes from the effects of large scale flow and mixing and it is for studies of continuous progressive evolution of atmospheric components that the use of a chamber as a well-mixed and closed batch reactor is suited. An additional important advantage of chamber studies is the ability to use relatively representative reactant concentrations and to study processes taking place over realistic atmospheric timescales. The reduced surface area to volume ratios with increasing reactor size enables this. A typical atmospherically relevant aerosol evolution timescale might be of the order of several hours to days. Therefore, chamber studies investigating the microphysical and chemical evolution of entire aerosol populations over realistic atmospheric timescales are possible. The underlying rationale in designing the experimental programme for the chamber facility outlined in this paper is to replicate aspects of atmospheric aerosol systems and, by close coordination of the experimental design with detailed model investigation of the relevant processes, aim to capture and understand the characteristics of the system determining the radiative implications. Such chamber experiments may not provide detailed mechanistic understanding of all processes contributing to the aerosol development (though extension of probing techniques is always possible), but those most pertinent to their atmospheric significance may be investigated in sufficient detail to test process model description. Throughout this proceedings issue a broad range of photochemical gaseous and aerosol chambers is described. The development of one of the most mature and largest photochemical chamber facilities is reported on http://pah.cert.ucr.edu/~carter/epacham/. It has been predominantly used to develop and validate the SAPRC-99 oxidation mechanism and is reported in Carter et al. (2003). It has not, however, been used extensively to investigate aerosol processes. One of the largest and most well appointed chambers used to investigate aerosol processes is the Caltech chamber (Cocker et al., 2001). It has been used to investigate the formation of aerosols from the photooxidation of diiodomethane (CH2I2) (Jimenez et al., 2003) and to develop and validate a model of secondary organic compound partitioning (Griffin et al., 2003) based on an extensive programme of aerosol formation experiments. Studies in the Kamens group facility at the University of North Carolina (Jang et al., 2003) have resulted in the discovery of condensed phase reactions leading to secondary organic aerosol formation and stabilisation. This concept has been further developed in the findings of Limbeck et al. (2003) showing acid catalysed aerosol formation from isoprene and Kalberer et al. (2004), where the condensed phase products after aerosol formation were shown to increase in molecular mass in the oxidation of 1,3,5-trimethylbenzene. The latter investigation was conducted in one of Europe’s largest indoor chamber facilities at the Paul Scherrer Institute in Zurich. There are other facilities in Europe where a range of aerosol investigations have been conducted. These range from the indoor chamber at Jülich and the EUPHORE chamber at Valencia to a range of smaller chambers. Most have some instrumentation for aerosol sizing and composition measurements, and a range of programmes, mostly limited to SOA formation studies in biogenically and anthropogenically

Field Measurement and Modelling Motivated Design of a Programme

51

influenced air masses (OSOA, EXACT, IALSI) have been conducted in these chambers. A mechanistic understanding of the formation of aerosols from photochemically formed gas phase precursors has been a priority objective in these studies, developing aerosol production potentials and yields in a range of systems, to provide an experimentally-constrained means by which secondary organic gas-to-particle conversion can be incorporated into useable models. However, many of the recent studies indicate the importance of secondary condensed phase reactions and complicated (yet not necessarily atmospherically representative) aerosol formation behaviour in the presence of high concentrations of mixed gas phase parent hydrocarbons. In addition, theories describing aerosol formation in chambers are not directly comparable to models including aerosol dynamics since the initial stage of nucleation is not considered, little molecular cluster and ultrafine aerosol number information is available and unrealistically high reactant concentrations are frequently used. For this reason, a primary difference between the studies outlined in this paper and the frequent SOA yield type experiments carried out in other chambers is that the experimental programme is designed to investigate aerosol evolution and growth, but will not consider formation. This avoids mechanistic interpretation difficulties of the formation process in terms of coupled chemistry and microphysics, but obviously provides no insight into nucleation and formation mechanisms. The experiment design thereby focuses only those aerosol process aspects directly accessible to description by relatively straightforward models. Approach to Chamber Design The design of a new aerosol chamber has followed from the range of atmospheric questions to be addressed in the experimental programme. Using climatic implications as both initial driver and final goal, the variety of investigations was broadly defined. In addition, the currently available aerosol modelling tools and those under development largely dictate the scope of the processes to be investigated. This paper therefore aims to describe the scientific questions, the modelling tools describing these questions, the proposed experimental protocol and the chamber design in response to the requirements. Range of Atmospheric Questions to be Addressed Avoiding investigations into aerosol formation processes, there are several distinct types of experiments which may be used to explore various aspects of atmospherically important systems including. x Non-reactive equilibration and microphysical evolution. Conventionally, the atmospheric aerosol has frequently been investigated assuming equilibrium is instantaneously attained between all semi-volatile components. This is conceptually adequate when considering systems which are not thought to be subject to kinetic limitations such as imposed by gaseous or aqueous diffusion or interfacial transport. The most straightforward systems would be equilibration of a population of small single involatile salt particles under varying

Figure 1. Predicted hygroscopic growth factor of NaCl / (NH4)2SO4 aerosol at 75% RH as a function of particle size in metres.

52

G. McFiggans

humidity and temperature environments. Equally accessible to description by available modelling techniques would be consideration of equilibration of a similar population of mixed involatile salt particles in a range of composition mixing states. Even in such simple systems, there are interesting predictions which may be investigated, such as that illustrated in Figure 1. Accounting for the effects of curvature on composition, calculations show that one might expect particles below a certain size to remain undeliquesced at higher relative humidities than those greater than this size (using the model of Topping et al. (2005a). More complex situations arise when semi-volatile components are present and the gaseous concentrations of these components will be important. Extension of the size distribution to include larger particles will introduce mass transfer limitation due to diffusion to the larger particles. Mixed semi-volatile component particles in such distributions may exhibit disequilibrium compositions. Introduction of organic components into the particles, either in externallymixed distributions or internally-mixed with inorganic components, may also introduce interesting phenomena. There is increasing evidence that organic components suppress efflorescence with reducing humidity (Marcolli et al., 2004). In addition, it is thought that some organic components take a significant time to attain equilibrium with ambient water vapour (e.g. up to 10 hours for glutaric acid (Peng et al., 2001)). Such equilibrium and kinetic effects will have significant atmospheric implications. In contrast to, for example, electrodynamic balance studies, large chambers provide a means of investigating such effects on entire aerosol populations. In addition, they offer longer residence times than flow systems (such as conditioners in TDMA studies) and hence the ability to better probe the extent of disequilibrium. Simple investigation of the changing equilibrium composition may be interpreted using thermodynamic models whilst the evolution of aerosol populations in chamber experiments over atmospheric timescales requires coupled microphysical and thermodynamic models. x Physical and chemical evolution of target aerosol system in isolation and externallymixed with a different seed background aerosol population with gas phase photochemistry. It may be appreciated that the introduction of gas phase photochemistry into systems such as described above will introduce a range of complications of atmospheric importance. Photochemical production or removal of semi-volatile components will perturb equilibria and hence change the evolution of the aerosol population. Production of reactive condensable species may initiate aqueous chemical reaction, further perturbing the equilibria. The aqueous chemistry can, in turn, effect the gas phase chemistry initiating complex multiphase cycles. Even in simple inorganic systems, such cycling may not be straightforward, e.g. halogen cycling through sea salt aerosol (Sander and Crutzen, 1996; Vogt

Field Measurement and Modelling Motivated Design of a Programme

53

et al., 1999; McFiggans et al., 2000; 2002). Figure 2 demonstrates the evolution of gas phase chlorine species resulting from such seasalt activation. Both gas phase and size-resolved aqueous phase composition measurements will be required. Cycling in organic systems will be hugely complex, but are likely to occur in the atmosphere. Large chambers provide sufficiently long residence times to establish such cycles and monitor the changes in aerosol and gas phase composition and physical properties on realistic atmospheric timescales. x Reactive development of target aerosol system in isolation and externally-mixed with a different seed background population and constant / varying gas phase photooxidant environment. It may be envisaged that the properties of atmospheric aerosol will be significantly affected by oxidation. For example, it is known that the oxygen to carbon ratio in fine particulate organic material increases with time moving away from an urban source region as does the hygroscopicity of the particles (McFiggans et al., 2005). Figure 3 shows the changing representative mass fragmentation pattern from an aerosol mass spectrometer deployed in urban and rural locations. This picture is common to all mid-latitude northern hemisphere findings. The more remote distributions are always dominated by a large contribution from CO2+ (m/z = 44) whereas the urban distribution Figure 3. A comparison of the measured mass shows a characteristic common fragmentation pattern and mass loading distribution of hydrocarbon fragmentation. The aerosol organics in different photochemical regimes as aerosol population evolves in a defined by the CO and O3 levels (HC, LO – high CO, changing oxidising environment low O3 and vice versa), (from Alfarra et al., 2004). and both surface and bulk oxidative processes may be involved in the observed changes. It is well established that both coagulation and condensation contributes to microphysical aerosol population evolution, but in addition to emerging field evidence for high molecular weight organic compounds contributing to the organic aerosol loading in a range of environments (O’Dowd et al., 2004; McFiggans et al., 2005), an increasing body of evidence for condensed phase macromolecule formation (e.g. Kalberer et al., 2004; Baltensperger et al., 2005) and surface oxidation (e.g. Donaldson et al., 2005) is being provided by chamber and laboratory studies respectively. This indicates that condensed phase chemical reaction may provide a significant, if not dominant, contribution to aerosol composition evolution. A simplification of the real

G. McFiggans

54

atmosphere as may be provided in chamber experiments is required to investigate the competitive effect of condensed phase oxidation, condensation and coagulation. In each of the three broad classifications of investigation above, significant uncertainties exist in the quantification of the effects of the respective processes on the ability of the aerosol population to affect climate forcing. The specific systems to be studied are described in the methodology section below. Available Modelling Tools Thermodynamics: A model of multicomponent inorganic and organic aerosol thermodynamics, ADDEM, has recently been reported (Topping et al., 2005 a, b). Accounting for the effects of particle curvature on composition, the equilibrium phase state of all involatile and semi-volatile components can be predicted at any given gaseous partial pressure of the components. For example, the model can be used to calculate the equilibrium size of multicomponent aerosol at a given relative humidity. Microphysics – aerosol and cloud: A variety of microphysical models accounting for condensational growth, evaporation and coagulation of aerosol particles in the sub- and supersaturated regimes, and cloud droplets in supersaturation, are available to investigate the evolution of condensed material under conditions of varying temperature and humidity. Coupled photochemistry and aerosol: Relatively simple schemes incorporating oddoxygen, odd-hydrogen and odd-nitrogen photochemistry and the oxidation of CH4, lumped hydrocarbons and reduced sulphur compounds have been coupled to the aerosol microphysical model described above. The coupled model may be driven by 2- or 4-stream radiative transfer algorithms to replicate the photon flux at any atmospheric level. A version of the model exists, extending that of McFiggans et al. (2000, 2002), describing the reactive cycling of halogens through seasalt aerosol particles. This version includes pH-dependent aqueous halogen chemistry. Mixed-phase cloud processes: A variety of mixed-phase and ice cloud models exist, describing the homogeneous and heterogeneous formation of water droplets and ice crystals. The implications of aerosol particles on mixed-phase clouds may be evaluated if their ice nucleating properties are known. Specialised coupled models: There are a number of detailed process models which can be used to address various coupled systems. One of particular interest to the chamber studies is the coupled model of aerosol formation and evolution in the photochemical iodine oxide system. Atmospheric Requirements Pressure: The chamber is designed to investigate boundary layer processes and will therefore operate at ambient atmospheric pressure. That the chamber does not require pressurisation allows the use of a flexible FEP Teflon bag. Teflon film is gas permeable, but with advantages that generally compensate for the drawbacks. It is broadly transparent to radiation at boundary layer wavelengths with well characterised transmission characteristics (Cocker et al., 2001). In addition, despite documented wall losses and re-emission of reactive

Field Measurement and Modelling Motivated Design of a Programme

55

species, Teflon is largely inert. Cleaning and flushing procedures will allow satisfactory operation in experiments involving predominantly unreactive aerosol systems. Humidity: The atmosphere is moist. The planetary boundary layer water vapour mixing ratio is of the order of a percent and relative humidity will range from low tens of percent to supersaturation. During the frequent cycling of air through the turbulent cloud-free or cloud-capped boundary layer, the atmospheric aerosol may experience relative humidities ranging from below the efflorescence point of pure salt solutions to supersaturations at which all but the smallest particles will activate into cloud droplets. A high purity water humidification system, humidity control and water vapour measurement are required in any aerosol chamber charging system. Light: Although many dark processes affect the evolution of aerosol populations (e.g. coagulation, condensation, evaporation, surface and condensed phase oxidation by ozone and NO3 radicals), a coupled atmospheric gas / aerosol system evolves in fluctuating solar illumination and will be subject to a range of photochemical processes. To investigate their effects, a chamber must be equipped with controllable illumination. One solution is to locate the chamber outside. Avoiding problems with naturally varying light intensity and exposure to inclement weather, indoor chambers must be provided with artificial lights. Artificial lights do not simulate natural sunlight across the full ground level solar spectrum, therefore photolysis rates between species will vary differently under natural and artificial irradiation. This presents various problems with characterisation and fluctuating intensity with wavelength investigated in detail elsewhere (Carter et al., 1995). Such issues must be accounted for when interpreting results. Artificial lights do have the advantage of allowing reproducible photolytic conditions between experiments. It has therefore been decided that the current chamber will use an artificial light source with the spectral characteristics monitored frequently, relating the investigated processes to atmospheric conditions using an explicit radiative transfer algorithm. Temperature: The lower limit of the range of temperatures to be probed within the chamber experiments is provided by the phase changes of water and the upper limit by reasonably expected boundary layer maximum. It is not proposed to investigate fog or cloud processes within the bag, so no operation below the dewpoint temperature at a given water vapour mixing ratio is envisaged. Neither is it envisaged that the chamber will operate above around 30qC. The lighting system will provide a large heat flux and the initial air charge will be warm due to heating in the high flow rate pumping system (see below). There will be no requirement for additional heating in illuminated experiments. There will, however, be a significant requirement for cooling. The temperature response of the system must be rapid and the thermal inertia will be high due to the large volume of air to be cooled. Aerosols in the Atmosphere: Atmospheric aerosol populations are complex. They are the result of variable heterogeneous primary emission sources, modification by superimposed effects of coagulation, condensation, evaporation and cloud processing, of various homogeneous and heterogeneous secondary formation processes and of size-dependent composition. They comprise a few inorganic salts, supposedly well-characterised in terms of their physico-chemical properties; a range of unmodified and reacted crustal mineral materials; elemental carbon / soot from combustion processes and hundreds, if not thousands, of largely unidentified primary and secondary organic compounds. This complexity in the composition of the particulate matter is only one dimension in the challenges provided by studies of atmospheric aerosol. The particles range in size from a few nanometres in diameter to many tens or even hundreds of microns. These four or five decades in diameter correspond to around 12 to 15 orders of magnitude difference in mass between the smallest and largest

G. McFiggans

56

particles. It can be seen that one or two large particles can vastly dominate the mass of the many thousands or tens of thousands of smaller particles. Unfortunately, however, the climate forcing resulting from aerosols is more strongly a function of the number of particles than the mass. For example, reflectivity and persistence of clouds is dependent on cloud droplet number which in turn depends strongly on the aerosol number. It is the composition of particles around the threshold activable dry size (somewhere near 100 nm diameter) which determines cloud droplet number, not the average aerosol bulk composition. A further significant complication is provided by the mixing-state in terms of the chemical composition and in terms of the derived physical properties. Each of the many components in an aerosol population may exist in a given size range of particles; the properties of the population will be significantly different if the components coexist in the same particles (“internally mixed”) or are in separate particles within the size range – “externally mixed”. Evidence from instruments measuring the water affinity of a particular size range indicate a range of affinities for a given size range in most locations (Cubison et al., 2005a). This indicates a range of compositions for any given size. Taken together, this evidence implies that extreme care must be taken when claiming that studied aerosol systems are of atmospheric relevance. Appropriate systems must be probed in an appropriate manner to make any inferences about their potential radiative impacts. Representative Gas and Aerosol Concentrations: It is usual for laboratory experiments and even many chamber experiments to be conducted at elevated concentrations of both gaseous components and aerosol concentrations. Unlike gas phase reactions which tend to be of relatively low order, aerosol microphysical process dependencies range over very many orders in concentration. For example, the nucleation process in the iodine oxide system is at least 4th order in atomic iodine concentration. Coagulation processes are dependent on particle number concentration and will be in competition with condensation and evaporation in their contribution to aerosol population evolution, both of which are dependent on something similar to the surface area (a surface area modified for a transition regime diffusion correction is one way of considering this) rather than the number concentration. The concentration regime in terms of gaseous components and in terms of several moments of the aerosol distribution must therefore all be considered when designing coupled system experiments. Undoubtedly the simplest means to achieve this conceptually is to conduct all experiments at close to ambient conditions. This may, however, present very many practical difficulties. Practical Considerations in Chamber Design Available Instrumentation: One of the most obvious constraints on chamber design and experimental programme is the availability of instrumentation to probe the quantities of interest. As outlined above, it is frequently necessary to simultaneously monitor aerosol physical and chemical properties in addition to gas phase chemical composition. A wide range of instruments are available for deployment in ground and aircraft-based field measurement within the Atmospheric Sciences Group at the University of Manchester. These will be available for deployment in the chamber in scheduled campaign mode and otherwise periodically when not in field use. In addition, there are a range of core instruments available for all experiments. The instrumentation suite is broadly classified as follows: Aerosol size distribution determination: A Differential Mobility Particle Sizer (DMPS) system comprising two (one short and one medium) Vienna-type Differential

Field Measurement and Modelling Motivated Design of a Programme

57

Mobility Analysers (DMAs, Winkelmayer et al., 1991) coupled to TSI model 3025A and 3010 Condensation Particle Counters (CPCs) with sheath to sample flow controlled volumetrically to 10:1 is continuously available to provide size distributions from 3 to 800 nm diameter (dry or at known relative humidity). The following sizing instruments are available periodically. A Grimm model 1.109 optical particle counter provides optical particle size distributions from 0.25 to 32 µm at ambient relative humidity by measurement of backscatter light intensity and inversion by Mie theory. A TSI model 3321 Aerodynamic Particle Sizer (APS) provides ambient aerodynamic particle size distributions from 0.5 to 20 µm diameter. A TSI model 3081 Scanning Mobility Particle Sizer (SMPS) is available for aerosol sizing across one of two standard sub-micron ranges. In addition, a variety of CPCs with differing size-cuts can provide rapid nano-particle fluctuation measurements. Using combinations of the above sizing techniques alongside vacuum aerodynamic diameter measurement from the Aerosol Mass Spectrometer (se below), it is possible to obtain information about particle morphology and density (DeCarlo et al., 2004). Aerosol composition, derived property and mixing state determination: An Aerodyne Aerosol Mass Spectrometer (AMS, Jayne et al., 2000) provides quantitative sub-micron nonrefractory aerosol component mass distributions by vacuum aerodynamic size. The loadings are commonly reported as sulphate, nitrate, ammonium and total organic distributions. In addition, the AMS has been used to discriminate between the characteristic organic mass fragmentation patterns in different locations (e.g. Figure 3). In laboratory and chamber experiments with aerosol of known initial composition, the AMS can provide useful mechanistic process information (Alfarra et al., 2005). A Hygroscopicity Tandem Differential Mobility Analyser (HTDMA; Rader and McMurry, 1986) comprising a medium length Vienna-type DMA coupled in series to a humidifier, a second DMA and a TSI model 3010 CPC is used to provide measurements of the affinity of the particles for water. The first DMA selects a pseudo-monodisperse dry cut from an ambient population, the humidifier subjects the cut to a known controlled relative humidity and the final DMA scans across an appropriate size range to determine the hygroscopic diameter growth factor, GFD; the amount by which the particles of this size grow in response to the changing humidity. The true ambient aerosol GFD is retrieved by inverting the observed distribution using an optimal estimation method (OEM, Cubison et al., 2005b). The HTDMA may also be used to infer the aerosol population component mixing state. The range of GFD at a particular size indicates the degree to which the compositions at this size vary between the particles. Combinations of AMS and HTDMA analyses will not yield unique solutions to the mixing state in the ambient atmosphere, but in chamber experiments with known initial composition, combined with some limited single particle composition information from the AMS, will provide almost all required composition information. Gas phase component concentrations: A number of standard gas analysers are routinely available for determination of gaseous mixing ratio of CH4, CO, O3, NO, NO2 and SO2 at moderate atmospheric concentrations. GC determination of a limited number of NMHC species is also available for appropriate experiments. An Aerodyne Quantum Cascade Tunable Diode Laser Absorption Spectrometer (QC-TDLAS) is periodically available for NH3 measurements using the line at 967 cm-1 and for NO2 measurement at 1606 cm-1. There are several potential experiments which may benefit from measurement of other trace species. The Differential Optical Absorption Spectrometer (DOAS) described below may be used for determination of IO, OIO, BrO, NO3, HCHO, SO2 and other trace species. Probing the potential for aerosol processes to affect impact on radiation: A key feature and major objective of the chamber facility is to investigate the changing potential of the aerosol populations to affect radiative forcing as they evolve chemically and physically

58

G. McFiggans

and thus to improve quantification of the effects of competing aerosol processes on climate. To fully quantify all potential effects it is necessary to investigate direct scattering of short and longwave radiation and longwave absorption in addition to indirect forcing by warm cloud activation of cloud condensation nuclei (CCN) and heterogeneous nucleation of ice nuclei (IN) in mixed phase / cold clouds. A Droplet Measurement Technologies CCN spectrometer, capable of measuring at supersaturations as low as 0.07 % is continuously available. An ice nucleus (IN) counter, built to the design of Rogers et al. (2001), is available to monitor the ability of the aerosol to act as ice nuclei. A DOAS system, comprising a Xenon arc lamp, active White cell optics traversing the minor axis of chamber coupled by optical fibre to a Princeton spectrometer with CCD detector is under construction. The DOAS system is not able to fully quantify direct aerosol scattering and absorption but will provide aerosol optical depth which may be interpreted alongside the direct size distribution measurements. Figure 4 shows a schematic of the chamber layout. It can be seen that the instrumentation requiring sample extraction is split between local and remote. The local instrumentation is located in the chamber laboratory, either fully enclosed in the environmentally controlled area (such as DMPS and HTDMA) or nearby. These instruments are on short bespoke inlets to minimise losses of, for example, ultrafine particles or “sticky” gases such as NH3. The “remote” instrumentation is located in the main aerosol laboratory above the chamber and allows the interfacing of the chamber to those instruments which are normally deployed in field experiments. Characteristic Requirements Common to All Chambers: The control of the environment will limit the useful experiments which are possible within the chamber. The environmental control provides the primary infrastructural requirements for the experimental programme. Charge supply: gas cleaning and aerosol generation: In experiments concerned primarily with investigation of nonreactive aerosol equilibration and microphysical processes, the requirements for gas scrubbing are less rigorous than for photochemical process studies and investigation of aerosol formation from parent hydrocarbons. However, a generally clean air supply is still required, particularly

Field Measurement and Modelling Motivated Design of a Programme

59

for those experiments investigating condensed phase oxidation processes. Particle filtration will be critical, as will the aerosol generation. The flush – fill cycling and experimental charge preparation system is driven by an oil-free 3 kW rotary side-channel blower. This is connected via 50 mm I stainless steel tubing via a series of pneumatically-operated 2- and 3-way fullbore ball valves and a series of filters. Currently gas scrubbing incorporates high flow rate activated carbon (Sofnocarb) and ultrafine particle (HEPA) filtration. Three additional conditioners will be added to the clean air supply (Molecular sieve, Moleculite and Sofnolime) to dry, convert CO to CO2 and remove CO2. The pump and all valves are computer controlled to cycle through chamber filling and flushing at greater than 1 m3 min -1 to enable many cleaning cycles between experiments. A corona discharge ozone generator is provided in-line to scrub the chamber of reactive hydrocarbons during the cleaning cycles. The aerosol generation is provided by a number of laboratory spray and nebulisation rigs feeding a 200 litre stainless steel reservoir through any combination of four inlet ports. This allows generation of any combination of internal and external aerosol component mixture to be fed to the experimental charge. Trace gas doping of the chamber is also possible through the reservoir. The flow is switched through the reservoir after the last cleaning cycle when delivering the experimental charge. Illumination and light flux measurements: The chamber illumination is still under design. The entire chamber is housed within an extruded aluminium frame. This will be enclosed within an outer walled frame to which the lamps will be fitted. The primary consideration is that a representative atmospheric light intensity of an appropriate spectrum is provided. The means of delivery is currently unresolved but will either comprise xenon or argon arc lamps or a combination of blacklights and other strip lighting. Both have advantages and the reader is referred elsewhere for a discussion of their relative merits (Carter et al., 1995). In any case, it is necessary to determine the wavelength-resolved light transmission of the chamber film. In addition, preliminary characterisation of the actinic flux across the chamber will be carried out using a Bentham DTM 300 scanning spectroradiometer (most appropriate in the UV) and a diode array spectrometer (with broader wavelength range and capable of more rapid measurement). Both spectrometers are available and for characterisation, the chamber is equipped with an adapter for the inlet manifold, rails to traverse the chamber and a miniature 2S input optical assembly. Relative humidity and temperature control and measurement: The strong temperature and RH dependence of the properties of interest in the target aerosol systems require that these quantities are continuously monitored and controlled such that fluctuations within the chamber throughout the duration of each experiment do not introduce undue errors. A chilled mirror hygrometer reference is linked to a number of capacitive RH and T sensors to continuously monitor the chamber conditions at a number of locations. The water vapour content of the chamber requires a pure water supply linked to the charging facility through a commercial steam generator. The temperature (and hence RH) control system is under design in conjunction with the illumination. The temperature control is necessitated largely by the lamp cooling requirement and also by the great sensitivity of the partitioning of semi-volatile components to temperature fluctuations. The air temperature between the clad outer frame and the film will be controllable to between 10qC and 30qC with appropriate response times and to an acceptable tolerance, nominally around 1qC fluctuation immediately following the lights being turned on and 0.5qC temperature drift throughout the experiments. This requires significant air handling and heat exchange (of the order of 30 to 40 kW with 8 to 10 kW of blacklight illumination). The temperature control and water vapour supply provides the RH control.

G. McFiggans

60

Chamber Dimensions and geometry: A key consideration in chamber design is the minimisation of reactant / aerosol losses and hence the maximisation of experimental duration. A larger chamber provides smaller wall area per unit volume and therefore a reduction in the wall effects. The geometry minimising the surface area per volume is a sphere. For practical purposes and cost effectiveness a nearly cubic chamber of 18 m3 has been chosen. The volume of air in the chamber will reduce through an experiment due to the sample volume requirement of the instruments. The chamber is mounted on three pairs of rectangular extruded Figure 5. Current chamber infrastructure. aluminium frames. The central pair are fixed whilst the upper and lower pairs are counterweighted and are designed to move in response to the reducing volume throughout each experiment (and during each flush / fill cycle). The chamber expansion and contraction is synchronised to the valve and pump operation by computer control. Figure 5 shows the current status of the chamber. Experimental Programme A number of preliminary studies have been defined to explore the suitability of the chamber for the three basic types of investigation detailed previously. The studies have been designed to be conducted throughout the early evolution of the working chamber. Experiment 1 is predominantly driven by predictions from a single and mixed salt thermodynamic model and will overlap significantly with the characterisation of the chamber. Predictions of the hygroscopicity of single-salt inorganic aerosols suggest that their deliquescence RH increases with decreasing size due to the Kelvin effect. The implication is that smaller aerosols may remain dry at the same RH that larger aerosols of the same composition will be deliquesced. Whilst possible to probe this effect systematically using a HTDMA alone, the process would be laborious and error-prone; likewise for electrodynamic balance experiments. This experiment is arguably the most straightforward, but may have profound implications on the direct scattering of light and on the surface reactivity of, for example, marine aerosols. After the initial characterisation stage, when wall losses of single salt aerosols will be investigated over a range of constant humidities, experiments will be conducted by nebulising broad aerosol distributions from single salt solutions into the chamber below the efflorescence RH of the aerosols at the inlet temperature. The chamber temperature will then be lowered, the size and composition of the distribution monitored and extinction measured by DOAS as the RH increases. Predictions of the changes in the distribution will be used to examine whether the measured aerosol extinction changes are consistent with suppressed deliquescence with size. This experiment will ascertain whether it is necessary to capture solid precipitation and deliquescence variation as a function of aerosol size in numerical descriptions of aerosol in order to realistically describe their

Field Measurement and Modelling Motivated Design of a Programme

61

direct radiative importance. For single involatile salts there is expected to be no change in indirect scattering, since this will depend on activating particles on increasing RH above ice or water saturation. Experiment 2 extends experiment 1 to consider multiple salt aerosols and mixed inorganic / organic aerosols. It is uncertain how distributions of even simple multiple salt aerosols behave with varying humidity. Similar predictions for multiple salts using the model from Topping et al. (2005a) indicate complex precipitation of double salts and multiple hydrates leading to staged deliquescence which varies as a function of size. Only by coupling the multicomponent activity model to a consideration of multicomponent surface tensions and densities can such predictions be made. Repeating experiment 1 with a range of multicomponent salts will allow these predictions to be tested. The multicomponent inorganic model has been extended to consider organic components using a group contribution theory approach (UNIFAC) (Topping et al., 2005b). A series of experiments investigating broad distributions of mixed inorganic / organic aerosols will be used to investigate the affect of the organic components on efflorescence / deliquescence as a function of size. As in experiment 1, if the model predictions are confirmed, it will indicate that size dependence of the deliquescence of multiple salt and mixed inorganic / organic aerosols must be captured to predict their scattering. In addition, recent electrodynamic balance work suggests that efflorescence is suppressed with an increasing number of organic species (Marcolli et al., 2004). This will be directly tested in this experiment by generation of appropriate test aerosols. A further complication may arise in these test systems in that equilibration timescales for some organic species in changing humidity environments may be of the order of several hours (Peng et al., 2001). This cannot be tested in the limited residence time allowed in HTDMA systems and requires the extended residence time allowed by chambers such as the one proposed. A final potentially important possibility to be tested is how the effect on the aerosol to act as CCN is affected by differential evaporation of volatile components as a function of size due to the extended equilibration timescales in the presence of organics combined with differential deliquescence with size. It is the aim to compare the experimental results with forthcoming developments in the mixed inorganic / organic thermodynamic model. Experiment 3 is driven by observations of changing aerosol properties as air moves downwind of a pollution source. The experiment will investigate processes contributing to the ageing of aerosols in a polluted plume. It is widely recognised that the composition of aerosols in polluted plumes evolves as the plume mixes into the regional background. This change is associated with a modification to the aerosol physical properties and an increase in their affinity for water. Since their affinity for water determines the ability for the aerosols to directly scatter radiation and to activate into warm cloud droplets, it is important to investigate the processes leading to the composition changes in aerosols in a polluted plume. A systematic chamber study programme will be used to explore the likely processes leading to these changes, firstly in simple single component systems of proxies for atmospheric organic aerosols, then in systems using several representative organic compounds combined variously in external and internal mixtures with inorganic components. Therefore, unlike those investigated in many chamber experiments, the polluted aerosol proxies will not be generated by nucleation from parent hydrocarbon oxidation, but will be nebulised from solutions of representative condensed phase components measured in the atmosphere. The generated aerosols will be subjected to oxidising environments in the absence of significant concentrations of gaseous hydrocarbons. The aim of this experiment is to investigate whether either surface oxidation processes or condensed phase reactions contribute to the aerosol evolution, particularly in terms of changes in the hygroscopic properties, thus changing their

62

G. McFiggans

radiative forcing potential (detected by CCN spectrometer, IN counter and DOAS). Hygroscopicity and composition measurements (HTDMA and AMS) can be directly compared with predictions from the mixed inorganic / organic thermodynamic model described above. No complex gas phase hydrocarbon oxidation mechanism will be required, hence oxidant concentrations under these experiments will be reasonably easy to predict. It is the intention to parameterise aerosol property changes in terms of oxidant concentration for incorporation into a coupled photochemical microphysical model. This experiment contrasts with those previous chamber experiments investigating the evolution of SOA nucleated in the presence of parent hydrocarbons in that it uses aerosols generated from compounds representative of those found in ambient samples as the reactant aerosols. These compounds may not prove to be of comparable reactivity to ambient aerosols, but may provide alternative insights to reaction of aerosols generated in unrepresentative parent hydrocarbon concentrations. This will help move towards a model of thermodynamic multicomponent inorganic / organic aerosol evolution and a parameterisation for condensed phase oxidation in sectional aerosol model. Experiment 4 is driven by atmospheric observations and previous laboratory studies and will investigate the atmospheric implications of coastal new particles formed during daytime low tide. Exposure of Laminaria macroalgae to atmospheric and elevated ozone levels in the laboratory has demonstrated that this is the source of bursts of iodine-containing particles in coastal experiments (McFiggans et al., 2004). Whilst mechanistic uncertainties remain, the most important question concerns the atmospheric significance of the particle “burst” phenomenon. The macroalgal ozone exposure reactor will be used to generate high ultrafine aerosol concentrations, comparable to coastal bursts. These will be fed into the chamber charge and evolve for several hours in the gaseous environment from the reactor outflow, and also in the same gaseous environment but seeded with SO2 to replicate that available from marine DMS oxidation. The primary aim is to ascertain whether particles formed from macroalgal emissions grow to radiatively active sizes in atmospherically reasonable timescales. For comparison, particles generated from the photooxidation of molecular iodine and diiodomethane (CH2I2) will be investigated in the same manner. DOAS retrieval of molecular iodine, iodine monoxide, iodine dioxide and sulphur dioxide concentrations will ensure that concentrations are representative of the coastal environment and enable a model of iodine particle nucleation and growth (McFiggans, 2002) to be constrained and applied to the chamber system. This experiment will provide a measure of the potential for iodine particles emitted in the coastal environment to participate in radiative forcing. Current Status The previous section describes a medium and long-term experimental programme. The characterisation of the chamber is currently at an early stage. Preliminary characterisation of the air supply and aerosol generation system has been conducted and some simple experiments attempted to establish the ability of the chamber to retain aerosol. Table 1 shows the average lifetime of relatively unreactive, involatile NaCl particles as a function of size measured during several 3 or 4-day experiments. Figure 6 shows an interesting phenomenon illustrating the forthcoming difficulties with chamber studies at low, atmospherically representative aerosol concentrations. The chamber was charged with an involatile NaCl seed aerosol in otherwise particle free scrubbed air. The illumination was with ambient laboratory lighting and daylight from the laboratory window.

Field Measurement and Modelling Motivated Design of a Programme

63

The decay of the number concentration with time at all sizes is evident over the first 12 hours or so in the DMPS measured size distribution. After this there is an obvious onset of nucleation and significant growth followed by decay in the aerosol number. The cycles of particle loss and fresh nucleation is repeated over several days, with no introduction of precursor gases. The initial concentration of NaCl particles is comparable to a low atmospheric loading and the nucleation is at a fairly low level. This illustrates the background level of aerosol dynamics which is currently experienced. This will improve with increased scrubbing and cleaning procedures.

Table 1. NaCl aerosol lifetime as a function of measured dry size.

Figure 6. Aerosol size distribution evolution during the first loss characterisation experiment.

Conclusions A chamber facility has been designed to investigate a range of atmospheric aerosol processes driven by their potential to affect radiative forcing. A preliminary experimental programme has been defined to complement an ongoing model development framework. The chamber infrastructure is partially constructed and undergoing preliminary characterisation. Illumination, temperature and humidity control systems are at various stages of design and construction and the experiments will follow the phased development of the chamber as appropriate. Acknowledgements The author would like to thank the UK Natural Environmental Research Council for support under grant NE/C000579/1 and would like to give thanks to the organisers of the Zakopane workshop on Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, NATO: EST.ARW 980164

64

G. McFiggans

References Alfarra, M. R., H. Coe, J. D. Allan, K. N. Bower, H. Boudries, M. R. Canagaratna, J. L. Jimenez, J. T. Jayne, A. A. Garforth, S. Li, and D. R. Worsnop; Organics by location, Atmos. Environ. 38 (2004) 5745-57. Alfarra, M. R., D. Paulsen, M. Gysel, A. A. Garforth, J. Dommen, A. S. H. Prévôt, D. R. Worsnop, U. Baltensperger, H. Coe; A Mass Spectrometric Study Of Secondary Organic Aerosols Formed From The Photooxidation Of Anthropogenic And Biogenic Precursors In A Reaction Chamber. submitted to Environ Sci Technol. (Oct. 2004). Baltensperger, U., M. Kalberer, J. Dommen, D. Paulsen, M. R. Alfarra, H. Coe, R. Fisseha, A. Gascho, M. Gysel, S. Nyeki, M. Sax, M. Steinbacher, A. S. H. Prevot, S. Sjögren, E. Weingartner and R. Zenobi; Secondary organic aerosols from anthropogenic and biogenic precursors, Faraday Disc. 130 (2005) doi: 10.1039/b417367h. Carter, W. P. L., D. Luo, I. L. Malkina, and J. A. Pierce, Environmental Chamber Studies of Atmospheric Reactivities of Volatile Organic Compounds. Effects of Varying Chamber and Light Source, Final report to National Renewable Energy Laboratory, Contract XZ- 2-12075, Coordinating Research Council, Inc., Project M-9, California Air Resources Board, Contract A032-0692, and South Coast Air Quality Management District, Contract C91323., March 26, (1995). Carter, W. P. L.; Evaluation of the SAPRC-99 Chemical Mechanism using new environmental chamber data, Presented at the Gordon Conference on Atmospheric Chemistry, Big Sky Resort, Montana, September 712, 2003 (available from: http://pah.cert.ucr.edu/~carter/bycarter.htm) Cocker, D. R. 3rd, R. C. Flagan, J. H. Seinfeld; State-of-the-art chamber facility for studying atmospheric aerosol chemistry, Environ Sci Technol. 35 (2001) 2594-2601. Cubison, M., M. Gysel, H. Coe and G. McFiggans; Are particles ever internally mixed? submitted to Geophys. Res. Letts. (2005) Cubison, M. and H. Coe; Use of an optimal estimation technique to retrieve ambient hygroscopicity distributions, submitted to J. Aerosol. Sci. (2005). Donaldson D. J., B. T. Mmereki, S. R. Chaudhuri, S. Handley and M. Oh; Uptake and reaction of atmospheric organic vapours on organic films, Faraday Disc. 130 (2005) doi: 10.1039/b418859d Griffin, R. J., D. Dabdub, J. H. Seinfeld; Development and initial evaluation of a dynamic species-resolved model for gas phase chemistry and size-resolved gas/particle partitioning associated with secondary organic aerosol formation, J. Geophys. Res. 110 (2005) doi:10.1029/2004JD005219 Jayne, J.T., D. C. Leard, X. F. Zhang, P. Davidovits, K. A. Smith, C. E. Kolb, and D. R. Worsnop; Development of an aerosol mass spectrometer for size and composition analysis of submicron particles, Aerosol Sci. Technol. 33 (2000) 49-70. Jang, M., N. M Czoschke, S. Lee and R. M. Kamens;. Heterogeneous Atmospheric Organic Aerosol Production by Inorganic Acid-Catalyzed Particle-Phase Reactions, Science 298 (2002) 814-817. Jimenez, J. L., R. Bahreini, D. R. Cocker, H. Zhuang, V. Varutbangkul, R. C. Flagan, J. H. Seinfeld, C. D. O'Dowd and T. Hoffmann; New Particle Formation from Photooxidation of Diiodomethane (CH2I2), J. Geophys. Res. 108, (2003) doi: 10.1029/2002JD002452. Kalberer, M., D. Paulsen, M. Sax, M. Steinbacher, J. Dommen, A. S. H. Prevot, R. Fisseha, E. Weingartner, V. Frankevich, R. Zenobi and U. Baltensperger; Identification of Polymers as Major Components of Atmospheric Organic Aerosols, Science 303 (2004) 1659-1662. Lewis, A. C., N. Carslaw, P. J. Marriott, R. M. Kinghorn, P. Morrison, A. L. Lee, K. D. Bartle, and M. J. Pilling, A larger pool of ozone-forming carbon compounds in urban areas, Nature, 405 (2000) 778-781. Limbeck A., M. Kulmala and H. Puxbaum; Secondary organic aerosol formation in the atmosphere via heterogeneous reaction of gaseous isoprene on acidic particles, Geophysical Research Letters 30 (2003) doi:10.1029/2003GL017738. Marcolli, C., B. P. Luo, and T. Peter; Mixing of the organic aerosol fractions: Liquids as the thermodynamically stable phases, J. Phys. Chem. A 108 (2004) 2216-2224. McFiggans, G., Plane, J. M. C., Allan, B. J., Carpenter, L. J., Coe, H. and O'Dowd, C; A modeling study of iodine chemistry in the marine boundary layer, J. Geophys. Res. 105 (2000) 14371-14385.

Field Measurement and Modelling Motivated Design of a Programme

65

McFiggans, G. Modelling the formation of aerosol due to iodine photochemistry – from the laboratory to the atmosphere?, Proceedings of the CERC 3 Young Chemists Workshop - Atmospheric Chemistry and Particulate Matter, Magleås, University of Copenhagen, Denmark, 3rd-6th June 2002 McFiggans, G., R. A. Cox, J. C. Mössinger, B. J. Allan and J. M. C. Plan; Roles of iodine and nitrogen in the release of active chlorine from marine aerosols, J. Geophys. Res. 107 (2002) doi:10.1029/2001JD000383. McFiggans, G., H. Coe, R. Burgess, J. Allan, M. Cubison, M. R. Alfarra, R. Saunders, A. Saiz-Lopez, J. M. C. Plane, D. Wevill, L. J. Carpenter, A. R. Rickard and P. S. Monks, Direct evidence for coastal iodine particles from Laminaria macroalgae – linkage to emissions of molecular iodine, Atmos. Chem. Phys. 4 (2004) 700-713. McFiggans, G., M. Alfarra, J. D. Allan, K. N. Bower, H. Coe, M. Cubison, D. O. Topping, P. I. Williams, S. Decesari, C. Facchini and S. Fuzzi; Simplification of the representation of the organic component of atmospheric particulates, Faraday Disc. 130 (2005) doi:10.1039/b419435g O'Dowd, C. D., M. C. Facchini, F. Cavalli, D. Ceburnis, M. Mircea, S. Decesari, S. Fuzzi, Y. J. Yoon, and J. P. Putaud; Biogenically-driven organic contribution to marine aerosol, Nature 431 (2004) doi:10.1038/nature02959. Peng, C. G., and C. K. Chan; The water cycles of water-soluble organic salts of atmospheric importance, Atmos. Environ. 35 (2001) 1183-1192. Rogers, D. C., DeMott, P. J., Kreidenweis, S. M., Chen, Y.; A Continuous-Flow Diffusion Chamber for Airborne Measurements of Ice Nuclei, J. Atmos. Oceanic Technol. 18 (2001) 725-741. Sander, R. and P. J. Crutzen; Model study indicating halogen activation and ozone destruction in polluted air masses transported to the sea, J. Geophys. Res. 101 (1996) 9121-9138. Rader, D. J. and P. H. McMurry; Application of the tandem differential mobility analyser for studies of droplet growth or evaporation, J. Aerosol Sci. 17 (1986) 771-787. Topping, D. O., G. McFiggans and H. Coe; A curved multi-component aerosol hygroscopicity model framework: 1 – Inorganics, Atmos. Chem. Phys. Discuss. 4 (2004) 8627-8676. Topping, D. O., G. McFiggans and H. Coe; A curved multi-component aerosol hygroscopicity model framework: 2 – Including organics, Atmos. Chem. Phys. Discuss. 4 (2004) 8677-8726. Vogt, R., R. Sander, R. von Glasow and P. J. Crutzen. Iodine chemistry and halogen activation in the marine boundary layer: A model study; J. Atmos. Chem. 32 (1999) 375-395. Winklmayr, W., G. P. Reischl, A. O. Lindner and A. Berner; A New Electromobility Spectrometer for the Measurement of Aerosol Size Distributions in the Size Range from 1 to 1000 nm, J. Aerosol. Sci. 22 (1991) 289-296. Winklmayr, W., G. P. Reischl, A. O. Lindner and A. Berner; A New Electromobility Spectrometer for the Measurement of Aerosol Size Distributions in the Size Range from 1 to 1000 nm, J. Aerosol. Sci. 22 (1991) 289-296.

Chamber Simulations of Cloud Chemistry: The AIDA Chamber Robert Wagner, Helmut Bunz, Claudia Linke, Ottmar Möhler, Karl-Heinz Naumann, Harald Saathoff, Martin Schnaiter, and Ulrich Schurath Forschungszentrum Karlsruhe, Institute of Meteorology and Climate Research (IMK-AAF), Karlsruhe, Germany Key Words: AIDA chamber, Ice nucleation, Polar stratospheric clouds, Soot aerosol, IR spectroscopy

Introduction Since its initial operation in 1997, the AIDA aerosol and cloud chamber of Forschungszentrum Karlsruhe (Aerosol Interactions and Dynamics in the Atmosphere) has been established as a unique experimental facility to study multi-phase processes over a wide range of atmospheric conditions. Research activities include heterogeneous chemistry on aerosols, hygroscopic and optical properties of complex aerosol particles, homogeneous freezing of supercooled solution droplets, heterogeneous freezing and cirrus cloud formation, as well as formation and characterisation of polar stratospheric cloud constituents.

Figure 1.

Schematic cross section of the AIDA facility. AIDA scientific instrumentation and aerosol generation devices are located on four separate floors surrounding the thermostated housing. 39 flanges with diameters from 0.01 to 1.1 m are located on the chamber walls, equipped with feed throughs for sensors, sampling tubes, and optical windows for various in situ spectroscopic measurements.

The AIDA chamber, a cylindrical aluminium vessel of 4 m diameter and 84.3 m3 volume, is located inside a large isolating container whose interior can be cooled to any temperature between ambient and 183 K (Figure 1). Temperature control is achieved by ventilating air around the aerosol vessel with upward directed flow rates of up to 30000 m3/h. Spatial and temporal temperature inhomogeneity is less than ± 0.5 K throughout the whole 67 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 67–82. © 2006 Springer. Printed in the Netherlands.

68

R. Wagner et al.

vessel. Cooling to 235 K is achieved by evaporation of a cooling liquid (R404A) inside two heat exchangers. For further cooling down to the minimum temperature of 183 K, liquid nitrogen is directly evaporated in the heat exchangers, leading to cooling rates of about 6 K/h at a liquid nitrogen consumption rate of approximately 0.6 m3/h. In addition to the broad accessible temperature range, the chamber can also be evacuated with two vacuum pumps from atmospheric pressure to a final pressure of about 0.01 hPa. This not only allows experiments to be performed at different pressure levels but also provides highly clean and aerosol free experimental conditions. Typically, background aerosol number concentrations are below 0.1 cm-3 after refilling the evacuated reaction vessel with particle free synthetic air. Further technical information about the thermostated housing, the vacuum pump, and gas supply system is available on the AIDA web page, http://imk-aida.fzk.de. In the following, a brief summary of major AIDA research activities is given, the advantages of carrying out these experiments in the AIDA chamber are highlighted, and recent journal publications ultilisating AIDA are referenced. Main components of the AIDA instrumentation will be described in the succeeding part of the article, including a comprehensive section about the application of FTIR spectroscopy to retrieve not only trace gas concentrations but also important aerosol parameters such as composition, phase, and size distribution. The remainder of this contribution will concentrate on ice nucleation experiments in the AIDA chamber. The operation of the AIDA vessel as a moderate expansion cloud chamber will be described and the data analysis of a typical ice nucleation experiment will be presented in detail. AIDA research activity since 1997 – an overview Early investigations with AIDA focussed on the heterogeneous reaction of ozone and other trace gases like NO2, HNO3, N2O5, NO3, and HO2 with soot aerosol (Kamm et al., 1999; Saathoff et al., 2001). The large volume of the AIDA chamber permitted experiments on time scales of several days, much longer than in previous laboratory studies. Therefore, it was possible to distinguish between reaction probabilities on typical tropospheric aerosol lifetimes and short-time saturation and passivation processes. In addition, the experiments were performed over a wide range of temperatures and relative humidities. Further studies placed emphasis on the detailed physical and chemical characterisation of airborne Diesel soot and spark generated soot particles, combining experimental and modelling approaches (Naumann, 2003; Saathoff et al., 2003a). Data analysis focussed e.g. on the optical properties of soot aerosol in the 230 – 1000 wavelength range (Schnaiter et al., 2003), as accurate optical data are urgently needed to assess the magnitude of the direct climatic impact of soot aerosol by absorption of solar radiation (Jacobson, 2001). The precise determination of absorption and scattering coefficients of pure soot particles prepared the ground for experiments dealing with the effect of the mixing state of soot with other aerosol particles on its optical properties (Saathoff et al., 2003b). Ageing processes like heterogeneous coagulation and condensation may change the mixing state of soot during its atmospheric lifetime of about one week, leading to soot particles coated with non-absorbing layers of organic and inorganic material. Theory predicts a strong increase in the specific absorption cross section upon coating, leading to enhancement factors of up to four, depending on layer thickness and core diameter (Fuller et al., 1999). To validate the model predictions, the optical properties of soot particles coated with different layers of organic material and sulphate were measured with the comprehensive optical instrumentation of the AIDA chamber. The organic coating was generated in situ by ozonolysis of D-pinene in the presence of soot aerosol. Due to the long residence time of the particles in the AIDA chamber, this coating procedure could be repeated

Chamber Simulations of Cloud Chemistry: The AIDA Chamber

69

several times, thereby gradually increasing the thickness of the non-absorbing organic layer. Data analysis clearly revealed an enhancement of the specific absorption coefficient of the soot particles upon coating by a factor of 1.8 at the maximum, thus confirming the previous model calculations. The formation of secondary organic aerosol (SOA) from the oxidation of biogenic compounds, used in the previous study to generate the organic coating of the soot particles, was also the central subject of independent AIDA experiments. Especially the ozonolysis of D-pinene, which is thought to be one of the major sources of SOA in the atmosphere, has been the focus of many previous laboratory studies. One major uncertainty, however, still remained the influence of temperature on the SOA yield. Therefore, due to the outstanding temperature control of the AIDA chamber, the yield of SOA material from the ozonolysis of D-pinene and limonene was recently measured in the temperature range from 253 K to 313 K. Data analysis, resulting in aerosol yields as function of temperature and aerosol mass, is in progress. First AIDA investigations dealing with multi-phase processes in the polar stratosphere started in 1998. From the beginning, these experiments were characterised by a fruitful cooperation with Max-Planck-Institute for Nuclear Physics, Division of Atmospheric Physics, Heidelberg. The research group in Heidelberg had developed a balloon-borne aerosol composition mass spectrometer (ACMS), allowing quantitative chemical analysis of polar stratospheric cloud particles (Schreiner et al., 2003). A similar spectrometer was used for simulation experiments in the AIDA chamber (Knopf et al., 2001). Initial studies focussed on the composition analysis of supercooled H2SO4/H2O solution droplets, acting as precursors for PSC particles (Zink et al., 2002). After demonstrating the suitability of the AIDA instrumentation to analyse composition, size, and mass concentration of the solution droplets in the temperature range from 235 K to 185 K, subsequent experiments dealt with particle growth and homogeneous freezing of the H2SO4/H2O droplets under simulated lee-wave conditions (Möhler et al., 2003). Particle freezing was induced by adiabatic expansion experiments in the AIDA chamber, which will be described in detail later in this contribution. In addition to aerosol composition mass spectrometry, satellite observations in the mid-infrared proved to be a valuable tool to derive chemical composition, phase, and size distribution of stratospheric sulphuric acid aerosols and polar stratospheric clouds. The retrieval of important aerosol parameters relies on accurate refractive index data sets for the various aerosol constituents, inferred from several recent laboratory measurements and comprehensively tabulated in the HITRAN database of spectroscopic parameters (Rothman et al., 2003). A comparison between individual low-temperature refractive index data sets, however, revealed significant spectral differences in important wavenumber regimes. These discrepancies motivated another set of AIDA experiments to test the accuracy of published refractive indices for supercooled sulphuric and nitric acid droplets. Aerosol composition and mass concentration of the solution droplets which were measured independently by various techniques could be directly compared to the retrieval results from the analysis of midinfrared extinction spectra, thereby assessing the accuracy of the available data sets of refractive indices (Wagner et al., 2003). Ongoing work focuses on the optical constants of supercooled ternary H2SO4/H2O/HNO3 solution (STS) droplets. STS droplets form at temperatures below 195 K during the polar night when nitric acid and water condense onto background sulphuric acid aerosols. This formation mechanism was successfully mimicked in an AIDA experiment by injecting supercooled H2SO4/H2O droplets (acting as background aerosol) at T = 194 K, followed by the addition of ozone and nitrogen dioxide. Gas phase

70

R. Wagner et al.

reaction between O3 and NO2 yielded N2O5, which then hydrolyzed in the liquid phase forming ternary H2SO4/H2O/HNO3 solution droplets with gradually increasing nitric acid concentration. This experiment will also be used to determine the uptake coefficient of N2O5 on liquid sulphuric acid and STS droplets at stratospheric temperatures. Most recent AIDA measurements in the framework of polar stratospheric cloud research addressed the nucleation of solid nitric acid dihydrate (NAD) from sub-micron (< 0.5 Pm) supercooled binary HNO3/H2O and ternary H2SO4/H2O/HNO3 solution droplets. These experiments will be analysed with the help of a process model to distinguish between volumeand surface-induced nucleation modes. First model runs indicate that the volume-based nucleation model compares much better with our experimental data (e.g. number concentration of the nucleated NAD crystals) than the surface-based scheme. This finding is important in the context of recent laboratory measurements on ice nucleation, revealing that homogeneous freezing of supercooled water droplets is a volume-dominated process at least for droplet radii well above 1 Pm (Duft and Leisner, 2004). However, experimental evidence that surface nucleation will also be unimportant for much smaller droplet radii typical for aerosol droplets in the atmosphere was still lacking. Another interesting aspect from the NAD nucleation experiments in the AIDA chamber became obvious from the analysis of the midinfrared spectra of the nucleated NAD crystals. The infrared spectral signatures indicated the formation of clearly aspherical, predominantly oblate NAD crystals with aspect ratios exceeding five (Wagner et al., 2004). Such a high degree of particle asphericity was never observed before in laboratory studies on nucleation of airborne nitric acid hydrate crystals, a difference which may be attributed to the longer observation time in the AIDA chamber. Obviously, nucleated NAD crystals gradually adopt a high degree of particle asphericity when slowly growing by mass transfer from unfrozen solution droplets in a supersaturated environment over a long period of time, a process also supposed to take place in the polar stratosphere. Studies on the homogeneous freezing of supercooled solution droplets and the heterogeneous ice nucleation on the surface of solid aerosol particles like soot or mineral dust turned out to be another promising area of application of the AIDA chamber. These processes are thought to play a crucial role for the formation and radiative properties of cirrus clouds. Data analysis concentrates on deriving freezing thresholds and critical ice supersaturations for different types of aerosols, thereby providing accurate input data for numerical models simulating cirrus formation (Haag et al., 2003). This demands great skill to precisely determine the freezing onset in the ice nucleation experiments. Due to the extended AIDA instrumentation, at least three independent methods can be relied upon to measure the onset of ice formation: Detection of > 0.5 Pm ice crystals with an optical particle counter, appearance of the characteristic extinction features of ice in the infrared spectra, and increase in the depolarization of back-scattered laser light. As indicated in the introduction, an exemplary ice nucleation experiment will be discussed in a separate section below. The first set of AIDA experiments in the context of ice nucleation focussed on the homogeneous freezing of supercooled sulphuric acid droplets in the temperature range from 235 K to 185 K (Möhler et al., 2003). Measured freezing threshold relative humidities agreed well with a recent activity-based parameterisation for describing homogeneous nucleation of solution droplets (Koop et al., 2000). According to this theorem, only the water activity of the aerosol particles but not their specific chemical composition determines the critical ice supersaturation for homogeneous freezing. By titrating the sulphuric acid solution droplets with ammonia in situ in the AIDA chamber, pure ammonium sulphate aerosol particles can be

Chamber Simulations of Cloud Chemistry: The AIDA Chamber

71

produced and subsequently used to study their ice nucleation potential. Remarkably, preliminary results have revealed that homogeneous freezing threshold relative humidities for ammonium sulphate solution droplets do not fit the activity-based theorem, showing significantly lower critical ice supersaturations compared to supercooled sulphuric acid droplets (Mangold et al., 2004). At present, however, it cannot be excluded that in the course of the titration process of sulphuric acid by ammonia gas some partial crystallization has occurred, yielding solid ammonium sulphate or letovicite, (NH4)3H(SO4)2. These solids may then have acted as nuclei for the supercooled ammonium sulphate solution droplets in the freezing experiments, reducing the ice supersaturation threshold compared to homogeneous freezing. Future AIDA freezing experiments are being considered to clarify this issue, thereby being careful to avoid any crystallization of solids in the supercooled ammonium sulphate solution droplets. Further AIDA ice nucleation studies addressed the ice nucleation efficiency of pure soot particles as well as soot particles coated with sulphuric acid (Möhler et al., 2004a). This work was part of a project investigating the impact of particles from aircraft engines on contrails, cirrus clouds, and climate. The results of these freezing experiments demonstrate that the mixing state of soot and sulphuric acid aerosol particles crucially affects the freezing threshold. Pure soot particles may act as efficient deposition ice nuclei in the temperature range from 235 K to 185 K, whereas coating with sulphuric acid significantly increases the critical ice saturation ratios, albeit still lying below the ice supersaturation thresholds for homogeneous freezing of pure sulphuric acid solution droplets. Very recent AIDA experiments have revealed the extraordinary properties of mineral dust aerosol particles in heterogeneous ice nucleation, acting as highly efficient deposition ice nuclei at any temperature between 195 K and 240 K (Möhler et al., 2004b).

T, p

Figure 2 .

Schematic view through the AIDA chamber, showing the most important devices for aerosol generation as well as trace gas and aerosol measurements. Note the mixing fan at the bottom of the vessel for achieving homogeneous temperature and humidity conditions inside the AIDA chamber.

R. Wagner et al.

72 AIDA instrumentation

Major components of the instrumentation of the AIDA chamber are shown in Figure 2. For a detailed listing of the complete technical equipment for trace gas and aerosol measurements, the reader is referred once again to the AIDA web page, http://imk-aida.fzk.de. Aerosol generation and characterisation. A broad variety of aerosol generation techniques has been designed for producing fundamental aerosol types, including sulphuric acid, nitric acid, and ammonium sulphate solutions droplets, airborne soot and mineral dust particles, as well as internally mixed aerosol particles like soot coated with sulphuric acid. Here, we want to draw attention to some special aerosol generation devices, i.e. our instrumentation for producing soot particles of different morphology and chemical composition as well as the apparatus for coating soot particles with sulphuric acid. Laboratory studies on airborne soot particles frequently employ an artificial soot sample from a commercial spark discharge generator (GfG1000) as surrogate for combustion soot. In the AIDA chamber, the properties of soot particles emitted from a VW Diesel engine have also been investigated, being available as part of the technical infrastructure of Research Centre Karlsruhe. It became obvious that the optical properties of airborne Diesel soot particles differ markedly from those of the synthetic soot sample, revealing, e.g., an increase in the specific extinction cross section by a factor of two at visible wavelengths (Schnaiter et al., 2003). More recent AIDA experiments introduced a new flame soot generator, the so-called combustion soot aerosol standard (CAST) burner. By slightly adjusting the CAST operation conditions, i.e. variations in the air/fuel(propane) ratio, soot aerosol samples with gradually varying organic carbon content could be generated. First results indicate that differences in the chemical composition significantly alter optical properties as well as the ice nucleation potential of the soot particles (Möhler et al., 2004b). Internal mixtures of soot or other aerosols like mineral dust with sulphuric acid can be generated in a homemade coating apparatus. Synthetic air which is saturated with sulphuric acid vapour by passing through a heated H2SO4 reservoir is mixed with the soot aerosol in a heated mixing chamber. Condensation of sulphuric acid vapour on the soot particles is achieved by passing the mixture through a temperature gradient tube. Size distribution measurements taken at the outlet of the tube section confirmed that condensation of sulphuric acid only occurs on the soot particles, i.e. without the formation of new particles. Higher temperatures of the sulphuric acid reservoir lead to an increase in the thickness of the sulphuric acid coating layer. Size distributions of the aerosol particles in the AIDA chamber are measured with a differential mobility analyser (DMA, 0.03 – 0.8 Pm range), connected to a condensation particle counter. As part of the AIDA experiments are dealing with partially volatile supercooled solution droplets like H2SO4/H2O, the DMA system was modified for operation inside the cooled isolating housing of the chamber (Seifert et al., 2004). This allows for measurements at stratospheric temperatures and pressures (183 K to 298 K and 100 hPa to 1000 hPa), thereby avoiding evaporation of the aerosol particles. Number concentrations and optical diameters of ice crystals in the 0.5 Pm to 20 Pm range are measured with two optical particle counters (PCS2000 and WELAS, Palas), mounted below the aerosol vessel and using vertical sampling tubes to minimise sampling losses.

Chamber Simulations of Cloud Chemistry: The AIDA Chamber

73

Humidity measurements Our reference instrument for measurements of water vapour in AIDA is a fast highprecision chilled-mirror hygrometer (MBW, model 373). For ice nucleation studies, however, special instrumentation is needed to detect the rapid changes in relative humidity during fast adiabatic expansion experiments in the AIDA chamber at very low temperatures and correspondingly low water vapour mixing ratios. A time resolution of 1 - 2 s is required to accurately determine threshold relative humidities of ice nucleation, which is much faster than the response time of the MBW under these extreme conditions. Thus, ice nucleation studies are performed in close cooperation with the ICG I group of Forschungszentrum Jülich, who contribute their Lyman-D hygrometer FISH which measures water vapour mixing ratios with a time resolution of about 1 s and a sensitivity in the order of 0.2 ppm (Zöger et al., 1999). Note that the FISH instrument is operated in an ex situ mode by sampling AIDA air through a heated inlet tube. As a consequence, the water mixing ratio measured by the FISH corresponds to the sum of interstitial (gas phase) and evaporated particle water. A valuable complement to recording the total water mixing ratio is the in situ measurement of the gas phase water concentration by tuneable diode laser (TDL) absorption. The diode laser which can be tuned in the range 1370 ± 2 nm is located in the laboratory area outside the cold AIDA containment. The laser beam is coupled via an optical fibre to an internal White-type multiple reflection cell, currently allowing an optical path length of about 130 m. This affords a sensitivity of better than 0.1 ppm water vapour at ambient pressure. Optical instrumentation Apart from the TDL system, a large number of additional in situ and ex situ spectroscopic techniques are used to study light absorption by trace gas molecules as well as light extinction (absorption + scattering) by aerosol particles in the AIDA chamber over a broad wavelength regime. For example, aerosol extinction can be measured without any frequency gap from 200 nm to 16.7 Pm using independent UV/VIS, NIR, and FTIR spectrometers. Here, we want to briefly introduce the set-up of two selected instruments: (i) the in situ laser scattering unit which allows for precise determination of the onset time of ice formation in AIDA expansion experiments, and (ii) the FTIR spectrometer - the analysis of aerosol extinction spectra in the infrared will then be demonstrated in the next section. (i) In this set-up, vertical polarized laser light from an Argon ion laser at 488 nm is directed horizontally through the AIDA chamber. Scattered intensities are detected in the forward (4° detection angle) and polarization resolved in the backward direction (176°), which enables to calculate the depolarization of the back-scattered laser light with a time resolution of 1 s. An increase in depolarization during AIDA expansion experiments sensitively indicates the formation of ice crystals, the non-zero depolarization being a result of their aspherical particle habits. (ii) FTIR spectra of trace gases and aerosol particles are measured in situ with a second White-type multiple reflection cell, yielding a horizontal optical path length of about 250 m. Spectra are recorded from 800 to 6000 wavenumbers at a resolution of 0.125 cm-1 for the analysis of gas kinetic experiments as well as 4 cm-1 when dealing with particle spectra, which proved to be sufficient to resolve their generally broad-band extinction features, as shown in the next section.

74

R. Wagner et al.

Retrieval of aerosol parameters from FTIR extinction spectra – a short survey The quantitative analysis of aerosol extinction spectra is based on the Lambert Beer equation (1). The measured optical depth W Q j at a specific wavenumber Q j may be expressed as

W Q j  log

I Q j I 0 Q j

L N ¦ n  Di V  Di ,Q j ln10 i 1

j 1" M , (1)

where L denotes the optical path length, n(Di) the number concentration of particles in a particular size bin Di of width 'D, and V(Di,Q j ) the size bin–averaged extinction cross section, given by

V  Di ,Q j

1 'D

Di 

'D 2

³

V  D,Q j dD. (2)

'D Di  2

As input for the retrieval algorithms, a M x N matrix of extinction cross sections, with N representing the total number of particle size bins and M the number of wavenumber points, has to be calculated (Arnott et al., 1997; Zasetsky et al., 2002). Extinction spectra of spherical solution droplets like H2SO4/H2O are computed with Mie theory, whereas the T-matrix method is frequently applied to account for the aspherical habits of e.g. water ice and nitric acid hydrate crystals, albeit being restricted to rotationally symmetric aspherical particles like spheroids and cylinders (Mishchenko and Travis, 1998). Literature data of the temperatureand frequency-dependent complex refractive indices of the various aerosol particles are required as input for both the Mie and T-matrix calculations. Using standard optimisation techniques, the summed squared residuals between experimental and calculated spectra are minimised using the individual size bin populations n(Di) as fitting parameters. In the case of H2SO4/H2O, HNO3/H2O, or H2SO4/H2O/HNO3 solution droplets, the fitting algorithm may be extended to not only retrieve size distribution parameters but also the aerosol composition, expressed in terms of wt% H2SO4 and/or wt% HNO3. Additionally, the aspect ratio of e.g. prolate and oblate spheroids or cylinders can be introduced as optimisation parameter to estimate the degree of particle asphericity when analysing infrared spectra of water ice or nitric acid hydrate crystals. To demonstrate the quantitative analysis of aerosol extinction spectra, Figure 3 provides a compilation of selected FTIR extinction spectra of various aerosol types, recorded during recent AIDA measurement campaigns (Wagner et al., 2003). The left figure shows measured infrared spectra of supercooled H2SO4/H2O solution droplets at 233 K and HNO3/H2O solution droplets at 192 K in comparison with best fit results from Mie calculations using the size distribution and the aerosol composition as adjustable parameters. The synthetic spectra rely on optical constants for H2SO4/H2O and HNO3/H2O solution droplets inferred from aerosol infrared spectra (Niedziela et al., 1999; Norman et al., 1999). Note the good agreement between measured and calculated droplet spectra, which is a first indication of the high accuracy of the refractive indices used in the Mie calculations. Even more impressive is the excellent agreement between the aerosol parameters retrieved from the FTIR extinction spectra and those independently measured with the additional AIDA instrumentation (e.g. filter sampling and humidity measurements). From the sulphuric acid droplet spectrum an aerosol composition of 55 ± 2 wt% H2SO4 and a sulphate mass concentration (derived from the retrieved size distribution and the density of the solution

Chamber Simulations of Cloud Chemistry: The AIDA Chamber

75

droplets) of 500 ± 25 Pg/m3 was inferred. This compares well with the complementary analytical results, namely 56.0 ± 0.5 wt% H2SO4 and 470 ± 50 Pg/m3 sulphate mass concentration. An equally good agreement was observed for the HNO3/H2O system, pointing out how accurately important aerosol parameters can be retrieved from FTIR extinction spectra if precise refractive index data are available.

H2SO4/H2O, 232.8 K

0.10

measured Mie fit

0.08 0.06 0.04

optical depth

0.02

0.04 0.03

NAD spheroids, calculated (c) optical depth / a.u.

optical depth

0.12

HNO3/H2O, 191.9 K measured Mie fit

0.02

NAD spheres, calculated (b)

measured (a)

0.01 0.00 6000 5000 4000 3000 2000 1000 -1

wavenumber / cm

Figure 3 .

2000

1600 1200 -1 wavenumber / cm

800

Left side: Measured FTIR extinction spectra of supercooled H2SO4/H2O and HNO3/H2O solution droplets in comparison with best fit results from Mie calculations. Right side: Measured D-NAD extinction spectrum compared to calculated spectra using different assumptions about particle shape. See text for details.

The right panel of Figure 3 illustrates the influence of particle asphericity on band positions and intensities in the infrared spectrum of nitric acid dihydrate in the 2200 to 800 cm-1 regime. Spectrum (a) shows the infrared extinction of D-NAD crystals which were formed via homogeneous nucleation of supercooled HNO3/H2O solution droplets at a temperature of about 195 K in the AIDA chamber. The measured spectrum is compared with two synthetic D-NAD spectra; spectrum (b) was calculated using Mie theory for a particle diameter of 1 Pm, and spectrum (c) was calculated with the T-matrix code for randomly oriented oblate spheroids with an aspect ratio of 15 and an equal-volume sphere diameter of 1 Pm, using optical constants of D-NAD retrieved from thin-film measurements (Toon et al., 1994). It is obvious that our measured D-NAD spectrum much better compares with the Tmatrix calculation, thus indicating the formation of strongly aspherical oblate D-NAD crystals (Wagner et al., 2004). The application of FTIR extinction spectroscopy to study composition and phase changes during a typical AIDA ice nucleation experiment is shown in Figure 4. The left panel shows a series of FTIR extinction spectra recorded during an AIDA expansion experiment

R. Wagner et al.

76

studying the homogeneous freezing of H2SO4/H2O solution droplets at 210 K. As described in detail in the next section, freezing could be initiated by reducing the pressure in the AIDA vessel with a mechanical pump. This results in cooling rates of up to 4 K/min, whereby freezing of the supercooled sulphuric acid droplets occurs when a threshold relative humidity is exceeded. During the freezing experiment depicted in Figure 4, aerosol extinction spectra were recorded every 40 s with the FTIR spectrometer. The bottom spectrum shows the infrared extinction of the sulphuric acid aerosol at static temperature and pressure conditions before pumping. The broad extinction feature at about 3300 cm-1 is due to the OH stretching mode of liquid water, absorption bands of the sulphate and bisulphate ions are located between 1000 and 1500 cm-1. After expansion cooling is started, the increasing relative humidity first of all leads to a dilution of the H2SO4/H2O aerosol particles which absorb water vapour from the gas phase. This is nicely evidenced by the continuous intensity increase in the liquid water extinction band at 3300 cm-1. Later on, distinct spectral changes occur in the grey-coloured spectrum: the frequency shift of the OH stretching mode as well as the appearance of a new extinction band at about 800 cm-1 both clearly indicate the formation of ice crystals. This indicates that the threshold relative humidity where supercooled H2SO4/H2O solution droplets freeze was exceeded. Note that the two subsequent infrared spectra shown in Figure 4 (scaled by factors of 0.6 and 0.4) are then completely dominated by light absorption and scattering of the growing ice crystals.

0.3 optical depth

*0.6

c(H2O) / ppm

*0.4

60 55 50 45 40 35 30 25 20

c(H2O) / ppm

0.4

35 30 25 20 15 10 5 0

0.2 't = 40 s

0.1

0.0 6000 5000 4000 3000 2000 1000 -1

wavenumber / cm

Figure 4 .

FISH: Total water TDL: Water vapour

0

200

400

600

FISH minus TDL FTIR retrieval 0

200

400

600

time relative to onset of pumping / s

Left side: Monitoring the homogeneous freezing of supercooled H2SO4/H2O solution droplets with FTIR extinction spectroscopy. Right side: Comparison of independent water measurements in the AIDA chamber. See text for details.

Quantitative analysis of FTIR spectra series recorded during AIDA expansion cooling experiments allows for deriving time profiles of the number concentration of the ice crystals as well as the ice water content (IWC). This provides a unique possibility to validate the independent measurements of these quantities with the optical particle counter as well as the FISH and TDL instruments, as demonstrated on the right side of Figure 4. The upper panel

Chamber Simulations of Cloud Chemistry: The AIDA Chamber

77

depicts time profiles of total (i.e. gas + condensed phase) and interstitial water vapour, measured by the FISH and by the TDL systems, respectively, during a homogeneous freezing experiment with H2SO4/H2O solution droplets at 224 K. Time zero on the relative time scale indicates the start of pumping. As the H2SO4/H2O mass concentration, and thus the water content of the supercooled liquid aerosol particles was quite low in this experiment, total water and gas phase water mixing ratios are identical until the freezing onset at about t = 60 s. After this time, the difference between total and the interstitial water directly yields the ice water content, as depicted in the lower panel of Figure 4 (right side). This time profile is compared with the ice water content retrieved from the simultaneously recorded FTIR extinction spectra. In this approach, the ice crystals were modelled as cylinders with a fixed aspect ratio of D/L = 0.7 (D diameter, L length of the cylinders). The ice water volume mixing ratios were calculated from the retrieved size distributions of the ice crystals. The excellent agreement between both independent methods to estimate the ice water content underlines the high precision and accuracy of the water measurements in the AIDA chamber, which provide an excellent basis for ice nucleation studies over a very wide temperature range (Mangold et al., 2004). The AIDA facility as moderate expansion cloud chamber for ice nucleation studies As already noted in the introductory sections, AIDA ice nucleation studies primarily focus on a precise determination of the critical ice saturation ratio for relevant aerosol types like H2SO4/H2O solution droplets, mineral dust and soot particles, as well as internally mixed aerosol particles. The ice saturation ratio Sice(T) pw  T (3) pw,ice T is defined as the ratio of the actual water vapour pw(T) and the saturation water vapour pressure over ice pw,ice(T) at the same temperature. Ice nucleation only occurs when a critical threshold relative humidity with respect to ice (RHi) is exceeded, which may be significantly larger than 100%. For example, 160-170% RHi must be exceeded to initiate homogeneous freezing of supercooled sulphuric acid solution droplets at temperatures below 210 K. On the other hand, mineral dust particles act as deposition ice nuclei already at critical ice saturation ratios well below 1.2. In the atmosphere, supersaturations with respect to the ice phase are induced by adiabatic expansion cooling of rising air parcels. In the AIDA chamber, ice saturation ratios of > 1.6 are established by controlled pumping, typically from 1000 to 800 hPa. The mechanical pump can be operated at variable pumping speeds to establish cooling rates which vary by more than an order of magnitude. Figure 5 depicts temperature and pressure profiles for two different pumping scenarios. Using the maximum pumping speed (from 1000 to 800 hPa in less than 4 min) leads to a cooling rate of initially more than 4 K min-1. At slow pumping speeds, cooling rates less than 0.1 K min-1 can be controlled. Note that the cooling rate in AIDA approximately obeys the Poisson equation (4) only at the very beginning of an expansion cooling experiment:

Sice T

T

§ p· T0 ¨ ¸ © p0 ¹

R Cp

(4)

Later on the cooling rate continuously levels off due to the increasing difference between the gas temperature and the wall temperature, which remains almost constant throughout the expansion. This gives rise to an increasing heat flux from the walls into the

R. Wagner et al.

78

chamber air and limits the maximum supersaturation with respect to ice which can be achieved.

Pressure (hPa)

1 000

950

1 900 850

2

800

Temperature (K)

240 238

1

236 234

4 K/min

2

232

0.1 K/min

230

0

10

5

15

Time (min) Figure 5 .

Time profiles of total pressure (top panel) and mean gas temperature (bottom panel) during two AIDA expansion cooling experiments with largely different pumping speeds.

Expansion cooling experiments are started after coating the walls of the AIDA chamber with a thin ice layer, i.e. at an ice saturation ratio Sice close to unity. The coating is made by filling the chamber with humidified air at some higher temperature and subsequent cooling to a lower temperature. Due to several internal heat sources (mixing fan, heated sampling tubes) the mean gas temperature is always a few tenths of a Kelvin higher than the wall temperature under static conditions, which explains why Sice is slightly less than unity before an expansion is started. As already indicated in the introductory section, very low background aerosol number concentrations of < 0.1 cm-1 can be established by evacuating the chamber to < 0.01 hPa and performing several flushing cycles with particle-free synthetic air. Afterwards, the desired seed aerosol is added to the chamber and characterised by e.g. filter sampling and size distribution measurements. Upon the start of expansion cooling, the calculated saturation water vapour pressure over ice pw,ice(T) in Eq. (3) decreases exponentially with decreasing AIDA gas temperature whereas the actual water vapour pressure pw decreases at most linearly with the total pressure during pumping. This leads to a rapid increase in Sice with a rate dSice/dt of typically 0.3-0.4/min. Note that pw is directly measured in situ with the tuneable diode laser instrument. The rate dSice/dt during expansion cooling is further enhanced by evaporation of ice from the chamber walls as a result of the almost constant wall temperature during pumping. The onset of ice nucleation after exceeding the freezing threshold relative humidity as well as time profiles of the number concentration of the growing ice crystals can be detected by various techniques. In the following, we will demonstrate this interplay of the various AIDA instruments, taking as example two ice

Chamber Simulations of Cloud Chemistry: The AIDA Chamber

79

nucleation experiments with different flame soot aerosol probes from the CAST burner as ice nuclei (Figure 6). This comparison will also illustrate how sensitively the ice nucleation potential depends on the chemical composition of the seed aerosol (Möhler et al., 2004b). Using different air/fuel ratios, two distinctly different samples of CAST soot aerosol with 16% and 40% organic carbon (OC) content, respectively, were generated. The OC fraction was determined by thermographic analysis of the soot aerosols sampled on quartz filters, following a standard protocol as described in the literature (VDI 2465, Blatt 2). To study the influence of the OC fraction of the soot aerosol on its ice nucleation potential, two otherwise identical AIDA expansion experiments were performed using either the 16% OC or the 40% OC soot samples as seed aerosol. The pressure changes during pumping as well as the associated changes in gas temperature for the two experiments are shown in the upper panels of Figure 6 (solid lines: 16% OC, dotted lines: 40% OC). Note the excellent reproducibility of the profiles, whereby the ice nucleation properties of different aerosol types can be directly compared. For the same reason approximately the same soot aerosol number concentrations (about 1000 cm-3) were used in both experiments. Panel 3 shows the relative humidity with respect to ice (RHi) calculated using Eq. (3) from the measured gas temperature and water vapour concentration (circles: 16% OC, triangles: 40% OC). When the onset time of ice nucleation can be determined, the corresponding threshold relative humidity can be directly read off this graph. At first, we will focus on the expansion cooling experiment with the 16% OC soot sample. Here, three independent measurement techniques point to a sharp, well-defined ice nucleation onset after about 100 s, as indicated by the solid vertical line. The optical particle spectrometer (PCS2000) starts to detect a large number of particles larger than 0.5 Pm in diameter, providing clear evidence for the formation and growth of pristine ice crystals (panel 5). At the same time, the infrared spectral signatures also indicate the nucleation of ice particles. The ice crystal number concentrations retrieved from the FTIR spectra (shown as circles in panel 5) turn out to be in excellent agreement with the data from the optical particle counter. Furthermore, the growing ice crystals lead to an increasing back-scattered light intensity, as evidenced in panel 4. From panel 3, we can infer an ice saturation ratio of 1.45 at ice nucleation onset. After reaching a maximum value of about 1.50, Sice starts to decrease as a result of the increasing uptake of water vapour by the growing ice crystals due to their increasing surface area concentration. Ice particle measurements in the expansion experiment with 40% OC soot aerosol markedly differ from the 16% OC sample. Note that the optical particle spectrometer hardly detects any ice particles. Additionally, extinction signatures of ice are barely visible in the infrared spectra and there is only a weak intensity increase of the back-scattered laser light in course of the expansion. The number concentration of ice crystals is less than 10 cm-3, thus < 1% of the seed aerosol particles act as deposition ice nuclei. In contrast to the 16% OC experiment, no precise critical ice saturation ratio can be specified for the 40% OC soot sample. RHi continues to increase to 190% because very little water vapour is lost on the small surface area of the scarce ice crystals. In summary, the comparison of the two expansion experiments provides first evidence that a higher fraction of organic carbon notably suppresses the ice nucleation potential of flame soot particles.

R. Wagner et al.

80

Figure 6 .

Measured time profiles of pressure, gas temperature, relative humidity with respect to ice, back-scattered laser light intensity, as well as ice particle number concentration for two expansion cooling experiments with different flame soot aerosol samples from the CAST burner as seed aerosol (Möhler et al., 2004b). See text for details.

Concluding remarks Due to the availability of several excellent diagnostic techniques for comprehensive studies of the formation and growth of ice crystals, as exemplified in Figure 6, ice nucleation on various types of aerosol particles will continue to be a focus of AIDA research activities. Future studies are planned to systemically analyse the impact of the chemical composition of soot on its ice nucleation properties. Further experiments will address the impact of coatings on the ice nucleating efficiency of mineral dust particles as well as the role of environmental ice nucleating bacteria. The above diagnostics will soon be supplemented by a new optical device which is presently under construction in our laboratory: it will be designed to measure

Chamber Simulations of Cloud Chemistry: The AIDA Chamber

81

scattering phase functions of ice crystals which are simultaneously imaged by a CCD camera with microscope optics at 2 µm optical resolution. References Arnott, W.P., C. Schmitt, Y.G. Liu and J. Hallett; Droplet size spectra and water-vapor concentration of laboratory water clouds: Inversion of Fourier transform infrared (500-5000 cm-1) optical-depth measurement, Appl. Opt. 36 (1997) 5205-5216. Duft, D. and T. Leisner; Laboratory evidence for volume-dominated nucleation of ice in supercooled water microdroplets, Atmos. Chem. Phys. 4 (2004) 1997-2000. Fuller, K.A., W.C. Malm and S.M. Kreidenweis; Effects of mixing on extinction by carbonaceous particles, J. Geophys. Res. (Atmos.) 104 (1999) 15941-15954. Haag, W., B. Kärcher, S. Schaefers, O. Stetzer, O. Möhler, U. Schurath, M. Krämer and C. Schiller; Numerical simulations of homogeneous freezing processes in the aerosol chamber AIDA, Atmos. Chem. Phys. 3 (2003) 195-210. Jacobson, M.Z.; Strong radiative heating due to the mixing state of black carbon in atmospheric aerosols, Nature 409 (2001) 695-697. Kamm, S., O. Möhler, K.H. Naumann, H. Saathoff and U. Schurath; The heterogeneous reaction of ozone with soot aerosol, Atmos. Env. 33 (1999) 4651-4661. Knopf, D.A., P. Zink, J. Schreiner and K. Mauersberger; Calibration of an aerosol composition mass spectrometer with sulfuric acid water aerosol, Aerosol. Sci. Tech. 35 (2001) 924-928. Koop, T., B.P. Luo, A. Tsias and T. Peter; Water activity as the determinant for homogeneous ice nucleation in aqueous solutions, Nature 406 (2000) 611-614. Mangold, A., R. Wagner, H. Saathoff, U. Schurath, C. Giesemann, V. Ebert, M. Krämer and O. Möhler; Experimental investigation of ice nucleation by different types of aerosols in the aerosol chamber AIDA: implications to microphysics of cirrus clouds, Meteorol. Z. (2004) accepted for publication. Mishchenko, M.I. and L.D. Travis; Capabilities and limitations of a current FORTRAN implementation of the Tmatrix method for randomly oriented, rotationally symmetric scatterers, J. Quant. Spectrosc. Radiat. Transfer 60 (1998) 309-324. Möhler, O., O. Stetzer, S. Schaefers, C. Linke, M. Schnaiter, R. Tiede, H. Saathoff, M. Krämer, A. Mangold, P. Budz, P. Zink, J. Schreiner, K. Mauersberger, W. Haag, B. Kärcher and U. Schurath; Experimental investigation of homogeneous freezing of sulphuric acid particles in the aerosol chamber AIDA, Atmos. Chem. Phys. 3 (2003) 211-223. Möhler, O., S. Büttner, C. Linke, M. Schnaiter, H. Saathoff, O. Stetzer, R. Wagner, M. Krämer, A. Mangold, V. Ebert and U. Schurath; Effect of Sulphuric Acid Coating on Heterogeneous Ice Nucleation by Soot Aerosol Particles, J. Geophys. Res. (Atmos.) (2004a) submitted. Möhler, O., C. Linke, H. Saathoff, M. Schnaiter, R. Wagner and U. Schurath; Ice nucleation on flame soot aerosol of different organic carbon content, Meteorol. Z. (2004b) accepted for publication. Naumann, K.H.; COSIMA - a computer program simulating the dynamics of fractal aerosols, J. Aerosol Sci. 34 (2003) 1371-1397. Niedziela, R.F., M.L. Norman, C.L. DeForest, R.E. Miller and D.R. Worsnop; A temperature- and compositiondependent study of H2SO4 aerosol optical constants using Fourier transform and tunable diode laser infrared spectroscopy, J. Phys. Chem. A 103 (1999) 8030-8040. Norman, M.L., J. Qian, R.E. Miller and D.R. Worsnop; Infrared complex refractive indices of supercooled liquid HNO3/H2O aerosols, J. Geophys. Res. (Atmos.) 104 (1999) 30571-30584. Rothman, L.S., A. Barbe, D.C. Benner, L.R. Brown, C. Camy-Peyret, M.R. Carleer, K. Chance, C. Clerbaux, V. Dana, V.M. Devi, A. Fayt, J.M. Flaud, R.R. Gamache, A. Goldman, D. Jacquemart, K.W. Jucks, W.J. Lafferty, J.Y. Mandin, S.T. Massie, V. Nemtchinov, D.A. Newnham, A. Perrin, C.P. Rinsland, J. Schroeder, K.M. Smith, M.A.H. Smith, K. Tang, R.A. Toth, J. Vander Auwera, P. Varanasi and K. Yoshino; The HITRAN molecular spectroscopic database: edition of 2000 including updates through 2001, J. Quant. Spectrosc. Radiat. Transfer 82 (2003) 5-44.

82

R. Wagner et al.

Saathoff, H., K.H. Naumann, N. Riemer, S. Kamm, O. Möhler, U. Schurath, H. Vogel and B. Vogel; The loss of NO2, HNO3, NO3/N2O5, and HO2/HOONO2 on soot aerosol: A chamber and modeling study, Geophys. Res. Lett. 28 (2001) 1957-1960. Saathoff, H., O. Möhler, U. Schurath, S. Kamm, B. Dippel and D. Mihelcic; The AIDA soot aerosol characterisation campaign 1999, J. Aerosol Sci. 34 (2003a) 1277-1296. Saathoff, H., K.H. Naumann, M. Schnaiter, W. Schöck, O. Möhler, U. Schurath, E. Weingartner, M. Gysel and U. Baltensperger; Coating of soot and (NH4)2SO4 particles by ozonolysis products of alpha-pinene, J. Aerosol Sci. 34 (2003b) 1297-1321. Schnaiter, M., H. Horvath, O. Möhler, K.H. Naumann, H. Saathoff and O.W. Schöck; UV-VIS-NIR spectral optical properties of soot and soot-containing aerosols, J. Aerosol Sci. 34 (2003) 1421-1444. Schreiner, J., C. Voigt, C. Weisser, A. Kohlmann, K. Mauersberger, T. Deshler, C. Kröger, J. Rosen, N. Kjome, N. Larsen, A. Adriani, F. Cairo, G. Di Donfrancesco, J. Ovarlez, H. Ovarlez and A. Dörnbrack; Chemical, microphysical, and optical properties of polar stratospheric clouds, J. Geophys. Res. 108(D5), 8313 (2003) doi:10.1029/2001JD000825. Seifert, M., R. Tiede, M. Schnaiter, C. Linke, O. Möhler, U. Schurath and J. Ström; Operation and performance of a differential mobility particle sizer and a TSI 3010 condensation particle counter at stratospheric temperatures and pressures, J. Aerosol Sci. 35 (2004) 981-993. Toon, O.B., M.A. Tolbert, B.G. Koehler, A.M. Middlebrook and J. Jordan; Infrared optical constants of H2O ice, amorphous nitric acid solutions, and nitric acid hydrates, J. Geophys. Res. (Atmos.) 99 (1994) 2563125654. Wagner, R., A. Mangold, O. Möhler, H. Saathoff, M. Schnaiter and U. Schurath; A quantitative test of infrared optical constants for supercooled sulphuric and nitric acid droplet aerosols, Atmos. Chem. Phys. 3 (2003) 1147-1164. Wagner, R., O. Möhler, H. Saathoff, O. Stetzer and U. Schurath; Infrared spectrum of nitric acid dihydrate influence of particle shape, J. Phys. Chem. A (2004) submitted. Zasetsky, A.Y., J.J. Sloan, R. Escribano and D. Fernandez; A new method for the quantitative identification of the composition, size and density of stratospheric aerosols from high resolution IR satellite measurements, Geophys. Res. Lett. 29, 2071 (2002) doi:10.1029/2002GL015816. Zink, P., D.A. Knopf, J. Schreiner, K. Mauersberger, O. Möhler, H. Saathoff, M. Seifert, R. Tiede and U. Schurath; Cryo-chamber simulation of stratospheric H2SO4/H2O particles: Composition analysis and model comparison, Geophys. Res. Lett. 29 (2002) 46-49. Zöger, M., A. Afchine, N. Eicke, M.T. Gerhards, E. Klein, D.S. McKenna, U. Mörschel, U. Schmidt, V. Tan, F. Tuitjer, T. Woyke and C. Schiller; Fast in situ stratospheric hygrometers: A new family of balloonborne and airborne Lyman-D photofragment fluorescence hygrometers, J. Geophys. Res. (Atmos.) 104 (1999) 1807-1816.

New On-line Mass Spectrometer for Identification of Reaction Products in the Aqueous Phase: Application to the OH-oxidation of N-methylpyrrolidone under Atmospheric Conditions Laurent Poulain, Anne Monod, and Henri Wortham Laboratoire de Chimie et Environnement, CNRS FRE 2704, Université de Provence case 29, 3 place Victor Hugo, 13 331 Marseille cedex 03, France Key Words : Aqueous phase tropospheric simulation chamber, On-line mass spectrometry, N-methylpyrrolidone, Kinetic rate constants, OH-oxidation, Reaction products, Chemical mechanisms

Introduction The atmosphere is a complex medium where chemicals are released, dispersed by physical processes and oxidized by photochemical reactions initiated by solar radiation. One of the most efficient oxidants is the OH radical produced by complex photochemical processes (Atkinson, 2000). In the past, the studies of Volatile Organic Compounds (VOCs) were focused on the gas phase reactivity. In the 1980s, inter-relations between gas and aqueous phases in the troposphere started to be considered (Graedel and Weschler, 1981). It was recognized that the aqueous phase photochemistry of Water Soluble Organic Compounds (WSOCs) has an impact on the gas phase concentrations of key species such as OH, HO2 and O3 (Lelieveld and Crutzen, 1990; Monod and Carlier, 1999; Herrmann, 2003). More recently, the contribution of WSOC to aerosol hygroscopicity and their ability to act as cloud condensation nuclei was found (Gelencser et al., 2003; Claeys et al., 2004; Ervens et al., 2004). Cloud droplets, which contain numerous chemical species originating both from the gas phase and from the condensation nuclei, are a medium where complex chemistry and photochemistry occur. Several free radicals (e.g. OH, NO3, Cl, Cl2-, SO4-) are formed, which can react with dissolved organic matter. Among these radicals, one of the most efficient oxidizing species is the OH radical (Ervens, et al., 2003, Herrmann, 2003). N-methyl-pyrrolidone (NMP) is used in the industry as a solvent. Its atmospheric lifetime is moderate, 13 h, if [OH] = 106 molecules cm-3 (Aschmann and Atkinson, 1999). However, NMP is a highly soluble compound (KH = 6.4 104 M atm-1, Hine and Moorkerjee, 1975), thus it is likely to enter into tropospheric droplets. The present work was aimed at determining the OH-oxidation rate constant of NMP and the reaction products formed in the aqueous phase under tropospheric conditions. NMP is a medium sized molecule (Table 1), containing 5 carbon atoms, so its OH-oxidation may give rise to a large number of products. This study presents a new on-line technique suitable to identify as many reaction products as possible, during the reaction. Two different kinds of experiments were conducted to determine respectively the kinetics and the reaction products formed. Experimental Reagents. The chemicals used were: H2O2 (not stabilized, Fluka, 30% w); methyl ethyl ketone (MEK) (Fluka, more than 99.5%); methyl-iso-butylketone (MIBK) (Fluka, more than 99%); FeSO4 (Prolabo), acetonitrile (Acros, HPLC grade); methanol (Acros, HPLC grade); succinimide (Sigma Aldrich, 98%), methylamide (Sigma Aldrich, 40%); N-ethylacetamide (Sigma Aldrich, 99%); N-methylsuccinimide (Sigma Aldrich, 99%); N-hydroxymethylpyrrolidone (Sigma Aldrich); acetamide (Sigma Aldrich, 98%), formamide (Sigma 83 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 83–96. © 2006 Springer. Printed in the Netherlands.

L. Poulain et al.

84

Aldrich, 98%), N-ethylformamide (Sigma Aldrich, 99%), 2-pyrrolidone (Sigma Aldrich, 99%), dimethylacetamide (Sigma Aldrich, 99%). Solutions were prepared using purified water from a Millipore MilliQ system utilising the reverse osmosis, micro-filtration, nucleargrade deionization and activated carbon modules (the resistivity of the purified water was greater than 1.8u 107 : cm-1). Glass cleaning procedure. The work was performed with diluted solutions, in order to avoid any contamination. The reactor and all the sampling vessels were carefully cleaned according to the following procedure: 3 rinses with reverse osmosed water; 1 hour soaking in 2 % DECON detergent diluted with reverse osmosed water; 5 rinses with ultrapure MilliQ water; 2 hours soaking in 2 % HCl diluted with ultrapure MilliQ water; 5 rinses with ultrapure MilliQ water; 15 hours soaking in ultrapure MilliQ water; 3 rinses with ultrapure MilliQ water. Kinetic studies of OH-oxidation of NMP We have shown previously that OH radicals can be generated through the Fenton reaction (1) under specific experimental conditions (Monod, et al., 2005). This technique was used in the dark at pH = 2 for the kinetic study of the OH oxidation of NMP. Fe

2+

H+

+ H2O2

OH

+

Fe(OH)2+

(1)

The kinetic rate constant of the OH oxidation of NMP was determined using the relative kinetic method. The principle of this method is to measure the decay rate of the OHinduced oxidation of the reactant (NMP) relative to that of a reference compound (R) for which the OH oxidation rate constant is well known (reactions 2 and 3). In this study, two reference compounds were chosen: methylethylketone (MEK) and methyl-iso-butylketone (MIBK): NMP + OH

R

+

OH

kX

kR

products

(2)

products

(3)

where kNMP and kR are the rate constants of the OH oxidation of NMP and a reference compound, respectively. The corresponding kinetic equations can be written as follows: § >NMP @ 0 · ¸¸ ln¨¨ © >NMP @ t ¹

§ >R @ 0 · k NMP ¸¸ u ln¨¨ kR © >R @ t ¹

(4)

where [NMP]0, [R]0, [NMP]t, [R]t are respectively the concentrations of NMP and a reference § >NMP @ 0 · § >R @ 0 · ¸¸ versus ln¨¨ ¸¸ yields a linear curve with compound at times 0 and t. Plotting ln¨¨ © >NMP @ t ¹ © >R @ t ¹ a slope of kNMP / kR and an intercept of zero. To obtain kNMP r dkNMP, the median value and

New On-line Mass Spectrometer for Identification of Reaction Products

85

§ >NMP @ 0 · ¸ ln ¨¨ the standard deviation (r2V ) of the ratios © >NMP @ t ¸¹ were calculated and multiplied by kR. § >R @ 0 · ¸¸ ln ¨¨ © >R @ t ¹

To establish the Arrhenius expressions for NMP, the equations given in Monod, et al. (2005) were used for the reference compounds:

ln k MEK (T )

26.2 r 1.0  1600 r 300

(5)

ln k MIBK (T )

25.6 r 1.0  1200 r 300

(6)

T

T

The analyses of NMP were performed using a High Pressure Liquid Chromatograph (HPLC) equipped with a reversed phase C18 column (Alltima, Alltech) and a UV detector working at 212 nm. A binary methanol/water eluent was used (with a gradient from 20%/80% to 90%/10% in 10 min, then maintained so for 5 min), at 1 mL min-1. The analysis of MEK and MIBK was performed by derivatisation with 2,4-dinitrophenylhydrazine (2,4-DNPH), followed by HPLC analysis with a reversed phase C18 column (Alltima, Alltech) and a UV detector working at 360 nm. A tertiary eluent (CH3CN/H2O/CH3OH) was used (isocratic at 20%/40%/40% for 3 min, then a linear gradient to 5%/25%/70% in 10 min, then another linear gradient to 5%/15%/80% in 10 min, then maintained for 6 min). Each sample was prepared by removing an aliquot of 180 µL from the reactor and diluting it with 800 µL of 2,4-DNPH (0.36 mg L-1). These techniques produced sharp peaks in the chromatograms. Before each experiment, a calibration was carried out for each compound in a range covering the concentrations encountered during the experiments.

23

T (K) 276

22.6 22,6

298 308

ln (kOH)

23

322

22.2 22,2

k OH + NMP (10 9 M-1 s -1) relative to MEK relative to MIBK (triangles) (squares) 2.69 r 0.52 2.44 r 0.37 4.63 r 0.51 3.50 r 1.3 4.30 r 0.54

3.45 r 0.31 2.99 r 0.47 3.66 r 0.40

21.8 21,8

21,4 21.4 21 21 3.00 3,00

3.10 3,10

3.20 3,20

3.30 3,30

3.40 3,40

3.50 3,50

3.60 3,60

3.70 3,70

10- 3 / T (K-1 )

Figure 1.

Arrhenius plot and relative rate constants (inset) for the reaction of OH radical with NMP in the aqueous phase.

L. Poulain et al.

86

The kinetic studies were conducted at four temperatures covering the atmospheric conditions, from 276 to 322 K. The results are presented in Figure 1. Taking into account all the data presented, one obtains the Arrhenius expression (7):

( 24.4 r 1.1) 

ln(k OH )

(760 r 320) T

(7)

The results show that the aqueous phase OH oxidation of NMP is fast compared to that of other WSOC, and that temperature has little effect on the kinetics. Assuming that 15% of air contains clouds, and that [OHg] = 106 molec/cm3, [OHaq] = 10-13 M, the presence of clouds can reduce the atmospheric lifetime of NMP by 13 %, from 13 h to 11 h (for the estimation of global atmospheric lifetimes, see Monod et al. (2005). Thus, the aqueous phase OH oxidation has an important impact on the atmospheric behaviour of NMP. Because the kinetics of the reaction is fast, special attention has to be paid to the reaction products formed. Reaction products studies

To study the reaction products the OH oxidation of NMP in the aqueous phase, the continuous photolysis of H2O2 was used to produce OH radicals (reaction 8). H2 O2

+

hX

2 OH

(8)

The Fenton reaction (1) was not used here to avoid any artefacts due to reactions of FeII with any organic acids formed. The reaction took place in an aqueous-phase simulation chamber, which consisted of a 450 mL Pyrex vessel. Unbuffered aqueous solutions were continuously stirred and maintained at 298 r 0.2 K. The light source was a 300 W xenon arc lamp (see Monod, et al. (2005) for further details on the apparatus). The initial concentrations of NMP and H2O2 were 5×10-4 and 1×10-3 mol L-1, respectively. During the reaction, aliquots of the solution were sampled at regular time intervals with an automatic sampler, and stored at 0 °C for further analysis. HPLC analysis

The formation of N-methylsuccinimide (NMS), succinimide and 2-pyrrolidone (2P) (Table 1) was expected by Campbell and Striebig (1999) and Friesen et al. (1999) who performed OH-oxidation of NMP in alkaline solutions and by photocatalysis, respectively. We performed the HPLC analysis using the same C18 column as described above, and a UV detection at 202 nm. A binary eluent (80 %/20 % acetonitrile/water) was used at 1 mL min-1. The results showed that some of the expected reaction products were formed. Figure 2a shows the time profiles of NMP, NMS and succinimide, while Figure 2b shows the corresponding plots of product concentrations versus the concentration of NMP consumed. The latter plots clearly show that NMS was a primary reaction product, whereas succinimide was a secondary one. The carbon balance of NMS production is 55 ± 1 %. This indicates that other primary reaction products have to be formed. Figure 2 shows that one unidentified compound was a primary product. Furthermore, the HPLC chromatograms contained a lot of unidentified and unresolved peaks (Figure 3). Therefore, an on-line mass spectrometer was designed in order to identify more reaction products.

New On-line Mass Spectrometer for Identification of Reaction Products

5x10

-4

4x10

-4

3x10

-4

2x10

-4

1x10

-4

0 0 NMP

10000

20000

30000

40000

time (seconds) 2P Succinimide

NMS

50000

2x10

6

2x10

6

1x10

6

5x10

5

0 60000

unidentified compound

[NMS] and [Succinimide] (mol/L)

Fig 2 b

2x10

-4

2x10

-4

1x10

-4

5x10

-5

0 0

1x10 NMS

Figure 2.

-4

2x10

6

2x10

6

1x10

6

5x10

5

unidentified compound (peak areas)

concentration (mol/L)

Fig 2 a

unidentified compound (peak areas)

87

0 -4 -4 -4 -4 2x10 3x10 4x10 5x10 ' [NMP] (mol/L) succinimide unidentified compound

HPLC analysis: (a) time profiles of NMP, two identified reaction products, (NMS and succinimid) and one unidentified reaction product, during the OHoxidation of NMP in the aqueous phase; (b) concentrations of the products (or HPLC signal in arbitrary units, a. u.) as functions of the consumed NMP.

L. Poulain et al.

88

Figure 3.

Example of an HPLC chromatogram during the OH-oxidation of NMP in the aqueous phase at t = 5 h 46 min (20778 seconds).

Design of an on-line mass spectrometer To perform in situ analysis of the solution during the course of the reaction, the simulation chamber was directly connected to a mass spectrometer (MS). The solution was transported to the MS by an HPLC pump. The MS was equipped with an electrospray ionisation (ESI) unit and a triple quadrupole, thus allowing ESI-MS+, ESI-MS-, and ESI-MSMS+ analyses to be performed. The HPLC pump flow was adjusted to 0.1 mL min-1 in order to get a good compromise between maximum analytical sensitivity and minimum volume sampled from the simulation chamber. A 9 minutes delay between sampling from the chamber and the MS analysis was taken into account in the evaluation of results. Over 20 hours of a reaction run, the mass spectrometer was operated continuously, in 48 cycles. Each cycle was 25 minutes long and contained 3 successive modes: ESI-MS+, ESI-MS-MS+ and ESI-MS(Figure 4), in order to identify the time profiles of as many compounds as possible. Each mode was operated with a scan range of 20 – 1000 atomic mass units (amu). The stabilisation times mentioned in Figure 4 were taken into account in the analysis of results. t = 6 min

t=0

ESI-MS positive mode

t = 14 min

ESI-MS-MS positive mode

Argon pressure stabilisation in quadrupole 2

t = 25 min

ESI-MS negative mode

Pressure stabilisation in quadrupole 2

Figure 4. A 25 minute operation cycle of the mass spectrometer directly connected to the simulation chamber during the OH oxidation of NMP in the aqueous phase. The cycle was repeated 48 times, to cover the 20 hours of a reaction run.

New On-line Mass Spectrometer for Identification of Reaction Products

89

Results. The experimental protocol allowed us to obtain, for the first time (to our knowledge), the time profiles of a large number of reaction products (around 66 species). The ESI-MS+ scans revealed around 40 products formed, and the ESI-MS- scans around 26 products. Figure 5a (full scale) shows the decrease of the NMP concentration during the reaction, at m/z = 100 amu in the positive mode (henceforward noted as 100+ amu), while Figures 5a (zoom) and 5b (full scale and zoom) show the formation of several oxidation products. Some of the products were formed initially and then destroyed, certainly due to their reaction with OH radicals (for example m/z = 86+, 114+, 130+ and 115-). Up to 300 amu, in both modes, one can observe the formation of compounds those masses are regularly spaced by 'm/z) = 14 to 18 amu (Figure 5a and 5b). This can be the consequence of the increasing oxidation level of the NMP ring, and/or the consequence of the formation of hydrated reaction products. Although the scan was performed up to 1000 amu, no signal was observed above 300 amu. Excellent reproducibility of the mass spectra was observed in three experiments. Discussions on the mass spectra. The identification of the products was performed by comparison with standard solutions (when available) and using ESI-MS+, ESI-MS-, and/or ESI-MS-MS+ scans. The results are shown in Tables 1 and 2. NMP was detected at 100+ amu. Interestingly, its dimer was detected simultaneously at 200 amu. Tests with standard solutions of NMP showed that the dimer was formed in the electrospray unit of the mass spectrometer, and the time-profile of the MS signal at 200+ amu seems to confirm that this dimer was not formed in the photoreactor (Figures 5a and 6). +

The time-profile of the MS signal at 100+ amu (Figure 6) is different from the NMP signal obtained by HPLC (Figure 2). The HPLC shows an exponential decrease in NMP, whereas the MS signal at 100+ amu shows a significant shoulder between 8000 and 22000 seconds, which is also observed in ESI-MS-MS+ scans at 58 amu. The shoulder can be explained by formation of a product of the same mass as NMP. Succinimide produced during the reaction was identified at 100+ amu, and the ESI-MS-MS+ scans of standard solutions revealed the same daughter ions as for NMP, at 58 and 72 (major) amu (Table 1). Fortunately, tests with standard solutions of succinimide showed that this compound and its hydrate can be detected in ESI-MS- scans at m/z = 97- and m/z = 115- (Table 2), and thus can be distinguished from NMP (Figure 6). In the HPLC analysis, N-methylsuccinimide (NMS) was identified as one of the main primary reaction products (Figure 2). Tests with standard solutions showed that NMS is detected in ESI-MS+ scans at m/z = 114+, and its hydrated form is detected in ESI-MS- scans at m/z = 130-. However, an excessive carbon balance (§ 120 %) was obtained with the MS analysis at m/z = 130- (calibrated with standard solutions of NMS), in disagreement with the 55 ± 1 % obtained by HPLC. This can be due to the formation of another product of equal mass. 1-Formyl-2-pyrrolidone (FP) has exactly the same mass as NMS, and was previously identified as an important primary reaction product during the OH oxidation of NMP in the gas phase (Aschmann and Atkinson, 1999). The ESI-MS-MS+ scans of mass 114+ during the experiments show the formation of two daughter ions: one at m/z = 86, which corresponds to NMS, as verified with standard solutions, and one at 72. FP is not available commercially, so we could not perform separate tests for this compound. However, the major fragmentation pathway of FP should include the loss of •CH2-C•(O) and/or •CH2-CH2-CH2•, by analogy with NMP, succinimide and NMS. The fragmentation pathway of FP leads to a major daughter ion at m/z = 72+, thus confirming the formation of FP during the reaction. Furthermore, FP may

90

Figure 5.

L. Poulain et al.

Full scale and zoom 3D plots of the MS signal versus mass and time during the OH-oxidation of NMP in the aqueous phase (scan range 20 to 300 amu), (a) in the positive mode, and (b, facing page) in the negative mode. See Table 1 for the identification of the observed masses.

also be responsible for the unidentified peak in HPLC (Figure 3), which signal shows that it is a primary reaction product (Figure 2). It is known that hydroxyl compounds can be formed in the aqueous phase during OH oxidation processes (Von Sonntag and Schuchmann, 1997), so this pathway was explored. At m/z = 116+ amu, Figures 5a and 6 show the formation of an important reaction product, which can be 5-hydroxy-N-methylpyrrolidone (5-HNMP) and/or Nhydroxymethylpyrrolidone (NHMP) (Table 1). The formation of both compounds was confirmed by ESI-MS-MS+ scans, with two daughter ions produced and detected at m/z = 88 (5-HNMP) and at m/z = 73 (NHMP). This is in good agreement with the major fragmentation pathways of NMP, NMS and succinimide. Furthermore, this was confirmed by tests with standard solutions of NHMP, which also allowed us to observe its hydrated form at m/z = 132- (Table 2). The formation of 2-pyrrolidone (2P) was suggested by Campbell and Striebig (1999) and Friesen et al. (1999). We observed its formation at m/z = 86+ amu (Figures 5a and 6), which was confirmed by ESI-MS-MS+ scans (Table 1). The formation of smaller compounds such as methylamine, formamide, NMF, N-methyl-4-aminobutanoic acid (in ESI-MS+ scans), and butanoic acid (in ESI-MS- scans) was also observed (Tables 1 and 2).

New On-line Mass Spectrometer for Identification of Reaction Products

Figure 6.

91

Time-profiles of some ESI-MS+ and ESI-MS- scans during the OH-oxidation of NMP in the aqueous phase.

92

L. Poulain et al.

Reaction products identified in the positive mode, and unidentified reaction products for which the MS signal was high.

Table 1.

Name

Formula

Methylamine

H3C

(M+H)+ a 32

NH2

m/z (MS-MS)+ b

O

Formamide

H2N

N-methylformamide (NMF)

46

C H O

H N

H3C

60

C H

H3 C

Acetamide

C

NH2

60

O

86

69 - 70

88

57-58

O

H

N

2-pyrrolidone (2P)

CH3

N-ethylacetamide

H2 C

H3C

H N

C

O CH3 CH3

Dimethylacetamide

H3C

N

88

C O

CH3

N

N-methylpyrrolidone (NMP)

100 58 (major) and 72 200 (NMP dimer)

O

H

Succinimide

N

O

O

100

72 (major) and 58

114

86 (major)

114

72 (major) supposed

116

88 (major) and 58

116

73

CH3

O

N-methyl succinimide (NMS)

N

O

H

O

C

1-formyl-2-pyrrolidone (FP)

N

O

CH3

5-hydroxy-N-methylpyrrolidone (5-HNMP)

HO

N

O

HO

CH2

N-hydroxymethylpyrrolidone (NHMP)

N

O

O

N-methyl-4-aminobutanoic acid

118 H3C

N H

OH CH3

N

O

O

2-hydroxy-N-methylsuccinimide

130 HO

Unidentified compound a

+

132 b

+

ESI-MS scans at (M + H); ESI-MS-MS daughter ions.

New On-line Mass Spectrometer for Identification of Reaction Products Table 2.

93

Reaction products identified in the negative mode, and unidentified reaction products for which the MS signal was high. Name

m/z (M – H + (M – H) H2O)- b

Formula

-a

H

N

O

Succinimide

97 c

O

H HO

Hydrated Succinimide

N

115 c

O

HO O

116

N-methyl-4-aminobutanoic acid H3C

N H

OH

CH3

Hydrated N-methyl succinimide (NMS)

HO

N

O

130

HO HO HO

O

C

Hydrated 1-formyl-2pyrrolidone (FP)

H

N

C

H

N

O

130

OH

or

OH

CH2(OH)

Hydrated Nhydroxymethylpyrrolidone (NHMP)

N

132

OH

OH

CH3

Hydrated 5-hydroxy-Nmethylpyrrolidone (5-HNMP)

HO

N

OH

132

OH CH3

HO

Hydrated 2-hydroxy-Nmethylsuccinimide

N

CH3

O

N

O

OH

146 OH

HO HO

or

HO

O

87

Butanoic acid OH

Unidentified compounds a

43 101 148

ESI-MS- scan at (M - H); b ESI-MS- scans of hydrated molecules at (M–H), Succinimide was identified at (M - 2H)- and its hydrate was identified at (M – 2H + H2O) after verification with standard solutions.

c

L. Poulain et al.

94 Chemical mechanisms.

Based on the results, a possible mechanism is suggested here for the OH oxidation of NMP in the aqueous phase under tropospheric conditions. The reaction probably proceeds via three different pathways (Figure 7). - Pathway (a) proceeds via hydrogen atom abstraction by the OH radical from the methyl group of NMP. This leads to the formation of an alkyl radical, which then reacts with dissolved oxygen to form a peroxy radical. By analogy with the aqueous phase behaviour of other peroxy radicals (Von Sonntag and Schuchmann, 1997), it can self-react to form a tetroxide which rapidly decomposes, leading to the formation of FP and NHMP. - Pathway (b) proceeds via hydrogen atom abstraction by OH radicals, from the CH2 group adjacent to the amine group of NMP. This leads to the formation of an alkyl radical which reacts with dissolved O2 to form another peroxy radical. This peroxy radical can also self-react to form a tetroxide, which rapidly decomposes into NMS and 5-HNMP. From the HPLC analysis a value of 55 ± 1 % can be put on the importance of this pathway. - Pathway (c) is more speculative, but it has been previously suggested by Horikoshi, et al., 2001, who investigated the OH oxidation of 2P in the aqueous phase, in the presence of TiO2. This pathway may occur for NMP under real tropospheric conditions. It proceeds via a ring-opening mechanism, leading to the formation of N-methyl-4-aminobutanoic acid. The formation of smaller molecules can be explained by the subsequent reactions of the primary products of the pathways (a), (b) or (c), which may lead to ring-opening mechanisms. CH3

N O

a

+ OH

c

b

O x

CH2

CH3

N

C

•

N

O

O

Cx

H3C

N

•

+

H2O

OH

+ O2

+ O2

C

• OO

CH3

CH2

• N

OO

O

CH2OH

CH3

N

N

FP

Figure 7.

NHMP

O

N-methyl-4aminobutanoic acid

O

N

O

O

N C

CHO

O

H3C

H N

CH3

O

NMS

HO

N O

5-HNMP

Suggested mechanism for the OH oxidation of NMP in the aqueous phase.

New On-line Mass Spectrometer for Identification of Reaction Products

95

Summary and conclusions The reactivity of NMP towards OH radicals was studied in the aqueous phase, under tropospheric conditions. The kinetic results show that the OH oxidation of NMP is fast compared to that of other WSOC, and thus should induce modifications of the composition of water droplets, due to the reaction products formed. A new experimental technique was developed to study the aqueous phase OH oxidation of NMP. A mass spectrometer was coupled to an aqueous phase simulation chamber, thus providing an on-line analysis of the solution. The mass spectrometer was equipped with an electrospray ionisation (ESI) unit and a triple quadrupole, which allowed ESI-MS+, ESI-MS-, and ESI-MS-MS+ analysis. The results proved that this experimental technique is highly promising, as it allowed us to detect the formation of 66 reaction products, of which 24 were positively identified. Based on the results obtained, a chemical mechanism has been suggested for the OH oxidation of NMP in the aqueous phase. The developed equipment can be used to study other molecules and other reactions of atmospheric interest. Acknowledgements This study was funded by the European project MOST (EVK2-CT-2001-00114). The authors also thank Séverine Barale, Marie Bonnevialle, Leïla Debiesse, Claire Detruit for their kind help with the experiments and analysis by HPLC. References Aschmann S. M. and R. Atkinson; Atmospheric chemistry of 1-methyl-2-pyrrolidinone, Atmos. Environ., 33 (1999) 591-599. Atkinson R.; Atmospheric chemistry of VOCs and NOx, Atmos. Environ., 34 (2000) 2063-2101. Campbell H. L. and B. A. Striebig; Evaluation of N-Methylpyrrolidone and its oxidation products toxicity utilizing the Microtox assay, Environ. Sci. Technol., 33 (1999) 11, 1926-1930. Claeys M., W. Wang, A. C. Ion, I. Kourtchev, A. Gelencser and W. Maenhaut; Formation of secondary organic aerosols from isoprene and its gas-phase oxidation products through reaction with hydrogen peroxide, Atmos. Environ., 38 (2004) 25, 4093-4098. Ervens B., G. Feingold, S. L. Clegg and S. M. Kreidenweis; A modeling study of aqueous production of dicarboxylic acids: 2. implications for cloud microphysics., J. Geophys. Res. [Atmos.], 109 (2004) D15, D15206/15201-D15206/15212. Ervens B., S. Gligorovski and H. Herrmann; Temperature-dependent rate constants for hydroxyl radical reactions with organic compounds in aqueous phase, Phys. Chem. Chem. Phys., 5 (2003) 1811-1824. Friesen D. A., J. V. Headley and C. H. Langford; The photooxidative degradation of N-Methylpyrrolidone in the presence of Cs3PW12O40 and TiO2 colloid photocatalysts., Environ. Sci. Technol., 33 (1999) 18, 3193 3198. Gelencser A., A. Hoffer, G. Kiss, E. Tombacz, R. Kurdi and L. Bencze; In-situ Formation of Light-Absorbing Organic Matter in Cloud Water, J. Atmos. Chem., 45 (2003) 25-33. Graedel T. E. and C. J. Weschler; Chemistry within aqueous atmospheric aerosols and raindrops, Rev. Geophys. Space Phys. 19 (1981) 505-539. Herrmann H.; Kinetics of aqueous phase reactions relevant for atmospheric chemistry, Chem. Rev. 103 (2003) 4691-4716. Hine J. and P. K. Moorkerjee; Strucural effects on rate and equilibriums. XIX. Intrinsic hydrophobic character of organic compounds. Correlations in terms of structural contributions, J. Org. Chem. 40 (1975) 293 298.

96

L. Poulain et al.

Horikoshi S., H. Hidaka and N. Serpone; Photocatalyzed degradation of polymers in aqueous semiconductor suspensions: V. Photomineralization of lactam ring-pendant polyvinylpyrrolidone at titania/water interfaces, J. Photochem. Photobiol. A: Chem. 138 (2001) 69-77. Lelieveld J. and P. J. Crutzen; Influence of cloud photochemical processes on tropospheric ozone, Nature, 343 (1990) 227-233. Monod A. and P. Carlier; Impact of clouds on the tropospheric ozone budget: Direct effect of multiphase photochemistry of soluble organic compounds, Atmos. Environ., 33 (1999) 4431-4446. Monod A., L. Poulain, S. Grubert, D. Voisin and H. Wortham; Kinetics of OH-initiated oxidation of oxigenated organic compounds in the aqueous phase: new rate constants, structure-activity relationships and atmospheric implications, Atmos. Environ. (2005) in press. Von Sonntag C. and H.-P. Schuchmann; Peroxyl radicals in aqueous solution, in: Z. Alfassi (ed.), Peroxyl radicals. John Wiley & Sons, Chichester (1997) 173-234.

Dynamic Chamber System to Measure Gaseous Compounds Emissions and Atmospheric-Biospheric Interactions Viney P. Aneja1*, Jessica Blunden1, Candis S. Claiborn2, and Hugo H. Rogers3 1

Department of Marine, Earth, and Atmospheric Sciences, North Carolina State University , Raleigh, NC 27695-8208, U.S.A. 2 Department of Civil and Environmental Engineering, Washington State University, Pullman, Washington 99164-2910, U.S.A. 3 National Soil Dynamics Laboratory, ARS-USDA, Auburn, Alabama 36832, U.S.A.

Key Words: Dynamic chamber, Emissions, Ammonia, Nitrogen oxide, Hydrogen sulphide, Hydrogen peroxide

Abstract The dynamic flow-through chamber system has been developed in response to a need to measure emissions of nitrogen, sulphur, and carbon compounds for a variety of field applications. The cylindrical chamber system is constructed of chemically inert materials and internally lined with 5mil thick transparent fluorinated ethylene polypropylene (FEP) Teflon to reduce chemical reactions and build up of temperature inside the chamber. The chamber (diameter = 27cm, height = 42 cm, volume = 24.05 L) is designed with an open-ended bottom that can penetrate either soil or liquid to a depth of ~6-8 cm, thus creating a completely enclosed system. Carrier gas (e.g. compressed zero-grade air) is pumped at a constant flow rate (~2 to ~5 lpm), depending on the season. The air inside the chamber is well mixed by a variable-speed, motor-driven Teflon impeller (~40 to ~100 rpm). Many different laboratory and field experiments have been conducted using this dynamic chamber system. Oxides of nitrogen (NO, NO2, NOy) emissions have been measured from agricultural soils where nitrogen-rich fertilizers have been applied. Ammonia-nitrogen (NH3-N) and reduced organic sulphur compounds emissions have been measured using this same technique across a gas-liquid interface at swine waste treatment anaerobic storage lagoons, and agricultural fields. Similar chamber systems have also been deployed to measure uptake of nitrogen, sulphur, ozone, and hydrogen peroxide gases by crops and vegetation to examine atmospheric-biospheric interactions. Emissions measurements have been validated by a coupled gas-liquid transfer with chemical reaction model as well as a U.S. Environmental Protection Agency (EPA) WATER 9 model. Introduction The dynamic flow-through chamber system is a technique that has been developed and modified over the past 25 years in an effort to measure earth-atmosphere and water-atmosphere fluxes of various compounds including biogenic sulphur, oxides of nitrogen, ammonia, and methane. Determination of fluxes for these different gaseous compounds is desirable in order to permit accurate assessment of the relative roles of biogenic and anthropogenic sources in contributing to such phenomena as the atmospheric sulphate and nitrate burdens, acidity in

97 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 97–109. © 2006 Springer. Printed in the Netherlands.

98

V. P. Aneja et al

precipitation due to biogenic emissions, and contribution to aerosol formation through atmospheric photochemical processes in localized areas downwind of suspect sources. Similar systems have also been utilized to measure exposure of plants to various air pollutants and to quantify the uptake of trace gases by plants (e.g., Aneja, 1976; Rogers et al., 1979; Heck et al., 1987; Claiborn and Aneja, 1993). Hill et al. (1978) initially developed this technique to measure biogenic sulphur fluxes from a salt marsh in Long Island, New York. More recently, the dynamic flow-through chamber system has also been used to measure ammonia emissions from agricultural and non-agricultural soils (Roelle and Aneja, 2002), hog waste treatment lagoons (Aneja et al., 2000, 2001a,b), and agricultural crop soils over which liquid hog waste has been applied (Roelle and Aneja, 2002), and biogenic nitric oxide emissions from various types of agricultural soils (Li et al., 1999; Roelle et al., 2001) as well as municipal wastewater treatment plant bio-solids-amended soils (Tabachow et al., 2002). Currently, a study by the North Carolina State University Air Quality research group, in coordination with the U.S. Department of Agriculture National Research Initiative, is underway to measure simultaneous emissions of nitric oxide, ammonia and hydrogen sulphide from swine waste treatment lagoons using this same technique. For measuring emissions, the chamber technique has the important advantage of association of a particular emission site and its measurable array of physical, chemical, and microbiological properties with emissions of particular compounds or their reaction products. In addition, gas residence times in the chamber are on the order of minutes so that chemical transformations between emission and analysis may be minimized. The entire chamber system is small, lightweight and easily transportable, thus making a survey of trace gas flux very convenient. Finally, the system is relatively inexpensive and laboratory requirements are low. However, it should be noted that the chamber system cannot completely simulate the ambient environment. The chamber measures over a small area (requires extrapolation) and extrapolation to large scale may be a problem. For measuring fluxes to plants, the dynamic chamber is operated as a well-mixed chamber. Ideally, then, mixing is instantaneous, conditions are uniform throughout the chamber, and exit conditions are representative of internal, well-mixed conditions, thus simplifying calculations of fluxes and mathematical treatment of non-plant surface losses (Aneja, 1976; Claiborne and Aneja, 1993). Dynamic Flow-Through Chamber for Emissions at the Air-Water/Soil Interface In the dynamic emission flux chamber method, a continuously stirred tank reactor (CSTR; Levenspiel, 1962) with an open bottom is placed over an area of interest of soil, mud, or water (Hill et al., 1978). For flux measurements at the air-water interface, a floating platform is used to hold the chamber system above water. A hole is cut in the centre of the platform in which the chamber rests, allowing the bottom of the chamber to penetrate the water surface by 3-4 cm, thus forming a seal between the water surface and the air within the chamber and thereby providing a completely enclosed system in which to measure gaseous fluxes. Figure 1 depicts a schematic of a typical dynamic flow-through chamber system used to measure gaseous emissions from a swine waste treatment lagoon. For soil measurements, the chamber is placed onto a stainless steel ring

Dynamic Chamber System to Measure Gaseous Compounds Emissions

99

which is then inserted into the soil. Flux measurements are taken ~1 hour after the insertion of the ring into the soil to ensure steady-state conditions have been reached within the chamber. Over the years construction of the dynamic chamber has been modified. For studies conducted recently, the chamber is generally constructed as follows. The dynamic flow-through chamber system is cylindrical in nature and built from Plexiglas material. Chamber dimensions may vary but are generally about 23 cm inner diameter (i.d.) and 46 cm. in height. The entire closed system is lined on the inside with 2 mil fluorinated ethylene propylene (FEP) Teflon and stainless steel fittings in order to minimize chemical reactions with sample flow.

Meteorological Tower

Zero Air (no NHx , NOx, or Sx)

Mobile Laboratory

Continuous Duty Motor

Vent Line

Mass Flow Controller

23 cm

Outflow Impeller Stirrer

Inflow

46 cm Floating Platform

pH Probe Lagoon

Temperature Probe #1 Temperature Probe #2

Figure 1.

Schematic of dynamic flow-through chamber system configured to measure emissions from a swine waste treatment lagoon.

A sweep gas such as compressed zero-grade air is passed through a Teflon FEP sample line into the chamber. The in-flowing air is monitored by a mass flow controller and subsequently delivered into the chamber. The air inside the chamber is ideally well-mixed by a variable-speed motor-driven Teflon impeller stirrer (speeds generally range from ~40-100 rpm). The dynamic chamber system, with the continuous stirring provided by the impeller, meets the necessary criteria for performance as a continuously stirred tank reactor (CSTR). For performance as a CSTR, the chamber needs to be “ideally” mixed (Aneja, 1976). In ideal mixing, the composition of any elemental volume within the chamber is assumed to be the same as that of any other volume within the chamber. Tracer experiments (residence time distribution) have been used to test the flow and mixing characteristics of the system. The results of these mixing studies

V. P. Aneja et al

100

indicated that the dynamic chamber behaved as a “perfect” mixer with negligible stagnancy or channelling (Aneja et al., 2000). Johansson and Granat (1984) conducted research on pressure differences between the outside atmosphere and air within a chamber using a tilted water manometer, which indicated that pressure differences were below detectable limits (0.2 mm H2O). Arkinson (2003) conducted research on temperature difference between the outside atmosphere and air within the chamber. The results indicate that temperature differences are d 0.4°C. The out-flowing gas also flows through a Teflon FEP sample line and is directed into a temperature controlled mobile laboratory, which houses the analytical instruments and data acquisition system. Here, the sample is analyzed continuously for content of the gas of interest. A vent line is fitted to the exiting sample line to prevent pressurization and was periodically bubble tested to check for under pressurization and/or leaks in the enclosed system. Sample lines do not exceed 10 meters. Flux Calculations The following mass balance equation may be used for the dynamic flow-through chamber system and applied to any target gaseous substance of known concentration: dC dt where

§ LAw q · § q[Cair ] JA ·  ¸R  ¨ ¸  C ¨ V ¹ V¹ © V © V

C

concentration of gas inside the chamber (ppbV)

Cair

concentration of gas in carrier air (ppbV)

q

flow rate of compressed air through the chamber (lpm)

V

volume of the chamber (L)

A

emission surface area covered by chamber (m2)

Aw

inner surface area of the chamber of inner and upper wall

(1)

surfaces (m2) L

total loss of gas in the chamber per unit area (m min-1) due to reaction with inner and upper walls of the chamber

h

internal height of the chamber (cm)

J

emission flux per unit area (µg [gas] m-2 s-1)

R

gas phase reactions inside the chamber

Since zero-grade air is used as the carrier gas, Cair is equal to zero and gas phase reactions, R, are also assumed to be zero. Since the air inside the chamber is assumed to be well mixed by

Dynamic Chamber System to Measure Gaseous Compounds Emissions

101

the impeller stirrer, C is assumed to be constant within the chamber. At steady-state conditions, § dC · the change of concentration with respect to time ¨ ¸ is expected to be zero. Therefore equation © dt ¹ (1) can be simplified as:

J h

§ LAw q · Ceq¨  ¸ V¹ © V

(2)

The loss term, L, is determined experimentally while equilibrium-state gas concentration (Ceq), flow rate (q) and chamber dimensions (V and h) are all measured. Kaplan et al. (1988) has ª Ceq  C (t ) º devised a method for calculating loss term by calculating the slope of the plot of  ln « » ¬ Ceq  Co ¼ versus time (t). For this experiment, Co is the initial equilibrium state gaseous concentration measured by the chamber system at a constant flow rate (lpm). Ceq is the measured gaseous concentration at a second equilibrium state at an increased or reduced flow rate (lpm) into the chamber system. C(t) depicts the gaseous concentration at any time, t, during the transition between the first and second equilibrium states. L is determined by:

L

q ·§ V · § ¨ slope  ¸¨ ¸ V ¹© Aw ¹ ©

(3)

Dynamic Flow-Through Chamber for Measuring Trace Gas Uptake by Plants For measuring uptake of gaseous species by plants, and for conducting exposure experiments, it is important to characterize the conditions very near the plant surface under study; thus, assuring well-mixed conditions for such experiments becomes very important. Claiborn and Aneja (1993) constructed a chamber with a 1:1 diameter to height ratio, which is ideal for wellmixed conditions (Uhl and Gray, 1966). The chamber is constructed of a Teflon-coated strap iron frame lined internally with a 5 mil thick Teflon film to allow for maximum light, and even light and temperature distribution. Mixing is provided by a Teflon-coated impeller mounted at the top of the frame, and aided by 3 Teflon-coated baffles. The entire system operates under slight negative pressure, and it is important to assure that there are no leaks into the system, to allow for proper determination of the air flow rate through the system and the initial concentration of the trace gas of interest. The air flow rate through the chamber provides one to two chamber exchange rates per minute. Sample ports are located on the inlet and outlet Teflon pipes. Separate sample manifolds are provided for CO2/water vapour and the trace gas of interest. Air samples from the chamber inlet and outlet are alternated automatically to the analyzers so that concentration differences across the chamber are determined using the same analyzer. In order to determine trace gas uptake, wall losses to the empty chamber surfaces must first be determined. The fraction of the trace gas introduced into the chamber that is lost to the chamber walls during the chamber loss experiment (CL) is determined:

V. P. Aneja et al

102

CL

C Ao ,empty  C A,empty C Ao ,empty

(4)

where CA and CAo are the outlet and inlet concentrations, respectively, of trace gas species A, and the subscript empty indicates the concentrations measured for the empty chamber experiment. Total losses (i.e., to both plant surfaces and chamber surfaces) are then measured, and the fractional loss (TL) of the trace gas entering the chamber is similarly determined:

TL

C Ao ,exp  C A,exp

(5)

C Ao ,exp

where CA and CAo are, again, the outlet and inlet concentrations, respectively, of trace gas species A, and the subscript exp indicates the concentrations measured for the exposure chamber experiment. Treatment of Wall Losses For trace gases that are not very reactive, it is sufficient to subtract the wall losses from the empty chamber experiment from the total trace gas losses from the plant uptake experiments in order to determine the losses to plant surfaces, alone. In other words, BL = TL-CL

(6)

where BL is the fractional loss of the trace gas to plant biomass surfaces, alone. For very reactive and/or water-soluble species like hydrogen peroxide or nitric acid, however, a more rigorous method of correcting for wall losses must be employed (Claiborn and Aneja, 1993). By treating both wall losses and losses to plant surfaces as two parallel reactions, the fractional loss to the plant biomass surfaces, BL, is calculated from TL and CL,

BL

TL  CL 1  CL

(7)

From this equation, it is evident that for species for which the chamber losses are very low, it is acceptable to simply subtract the fractional loss to the walls (CL) from the fractional loss during the exposure chamber (TL). Finally, fluxes to plant surfaces (QA) are calculated from the fractional loss to the plant surfaces (BL), the chamber air flow rate (F), the inlet trace gas concentration, and the plant surface area (AL): QA

BL F C Ao,exp AL

(8)

Dynamic Chamber System to Measure Gaseous Compounds Emissions

103

Results The graphs in Figure 2 (a,b,c,d) provide results from some field studies undertaken to measure fluxes of selected S, N, and C compounds. Figure 2a (Aneja et al., 2001c) reports the log of hourly averaged ammonia-nitrogen flux from a swine waste treatment lagoon plotted against lagoon aqueous phase surface temperature during four different seasonal measurements. The graph shows an observed exponential (r2 = 0.78) relationship between NH3 and lagoon water temperature measured over the year. Figure 2b demonstrates the effect of soil temperature on NO from an agricultural (soybean) crop soil at four different sites. Figure 2c illustrates the temperature dependence of average H2S emission fluxes from a salt marsh. The emission rate appears to increase semi-logarithmically with increasing temperature within a temperature range of ~0qC to ~40°C. Finally, Figure 2d represents isoprene (hydrocarbon) concentration plotted against temperature during warm months (April through September). Please note that concentrations (ppbV), which are treated as a surrogate for emission (since the concentration measurements are in close proximity to the source), are reported rather than flux calculations. The graph depicts the linear relationship between the logarithm of isoprene concentration and temperature using data from four sites.

a)

(c)

Figure 2 (a,b,c,d).

(b)

(d)

Emissions of NH3 (Aneja et al., 2000), NO (Aneja et al., 2001c), Isoprene (Hagerman et al., 1997), and H2S (Hill et al., 1978), respectively, versus Temperature at various locations.

V. P. Aneja et al

104

The well-mixed chamber was used to measure uptake of gaseous hydrogen peroxide by spruce saplings (coniferous species). Hydrogen peroxide is a highly reactive and water-soluble atmospheric oxidant, for which wall losses are very significant (10-35% for 1-2 air exchanges per minute) and highly variable (standard deviation of 50% for daytime experiments) (Claiborn, 1991). A linear relationship between hydrogen peroxide flux and chamber concentration was apparent for the red spruce saplings (Figure 3). Note that there is no significant difference between daytime and night time fluxes. Average deposition velocities were calculated for these experiments, and ranged from 0.08 to 0.12 cm/s, and 0.08 – 0.12 cm/s for daytime and night time values, respectively. In contrast, similar uptake experiments conducted on bean plants (broad leaf plants) resulted in deposition velocities ranging from 0.8 to 1.4 and 0.5 to 0.6 for days and nights, respectively (Claiborn, 1991). H2O2 uptake by Red Spruce saplings 25

Day Night

Flux, x 10^15 mol/cm2-s

20

15

10

5

0 0

1

2

3

4

5

6

7

Chamber concentration, ppbv

Figure 3 .

Fluxes of gaseous H2O2 to needles of red spruce saplings as measured in wellmixed chamber experiments. Diamonds denote daytime fluxes and squares denote night time fluxes (Claiborn, 1991).

Data Comparisons As with any type of experimental procedure, it is desirable to compare data acquired in field studies with controlled laboratory studies and/or related models. Various comparisons with models, as well as other measurement techniques, have been made with data collected via the dynamic flow-through chamber system. Some results are presented in the following discussion. For model comparisons with the dynamic flow-through chamber system, Aneja et al., (2001d) developed a mass transport model based on the quiescent thin film concept (Danckwerts,

Dynamic Chamber System to Measure Gaseous Compounds Emissions

105

1970), which takes into account molecular diffusion and chemical reactions. Figure 4 shows seasonal variation comparison of ammonia emissions between modelling results and dynamic chamber system experiments for ammonia emissions from a swine waste treatment lagoon from a field study conducted by Aneja et al. (2000). The coupled mass transfer model simulations corroborated experimental results utilizing the dynamic chamber technique and validated the application to the flux equation. Also, Figure 5 shows how closely the recently developed WATER9 model, based on a set of air emission models that are documented in Air Emissions Models for Waste and Wastewater, U.S. Environmental Protection Agency, (U.S. EPA, 1994), compare with the same set of ammonia emissions data as discussed with the mass transport model previously. Tabachow et al. (2002) conducted and compared the results of nitric oxide (NO) emissions measurements for both in situ field and laboratory experiments from unamended and municipal wastewater treatment plant biosolids-amended soil. Based on matching soil temperatures (within 5ºC) and water filled pore space (WFPS) (± 5%), the ratio of volumetric soil water to total porosity of the soil, five of six scenarios showed no statistically significant difference in the NO flux measurements from the laboratory versus the field studies.

Figure 4 .

Seasonal variation comparison of ammonia emissions between modelling results and dynamic chamber system experiments.

Nitric oxide emissions from soils on a soybean field, measured by the dynamic chamber system (Li et al., 1999), were compared in Figure 6 with NO fluxes calculated at a 5 meter height utilizing the eddy-correlation method (Gao et al., 1996). In spite of differences in flux magnitudes (note the different scales used), the two show qualitatively similar variations and structure with time. The values of NO soil emissions are greater than the NO fluxes at 5 m, which implies that

V. P. Aneja et al

106

some of the NO might have been converted to nitrogen dioxide (NO2) through reactions with ozone and peroxy radicals by the time it reaches the 5-m height level from the source. In a coordinated research effort (Project NOVA), NO emissions were measured from fine sandy loam soil in a corn field using two different (dynamic and static) chamber techniques. One group from North Carolina State University used the dynamic chamber method while a group from NASA Langley Research Center used a closed box flux technique (static chamber), in which NO fluxes were calculated using the mixing ratio of NO (ppbv) versus time. (Anderson and Levine, 1987). Figure 7 depicts a difference plot of the composite averaged fluxes calculated by the dynamic chamber technique and the static chamber technique. Upon, statistical analysis, it was shown that there was no statistically significant difference between NO flux as measured using the two different chamber methods. Chamber experiments on uptake by plants provide useful information for the determination of deposition velocities, however, the deposition velocities over forest canopies will be different to those calculated in the chamber experiments. Deposition velocities take into account not only the resistance to mass transfer due to the plant itself, but also the resistance due to the forest canopy structure, and the turbulence over the canopy.

U.S. EPA WATER9 Model Ammonia Flux Measurements

Figure 5 .

Ammonia flux versus lagoon temperature comparison between WATER9 modelling results and dynamic chamber experimental calculated results.

Dynamic Chamber System to Measure Gaseous Compounds Emissions

107

Figure 6 .

Inter-comparison of NO fluxes at 5 meter height using eddy-correlation technique with NO soil emissions using dynamic chamber technique.

Figure 7.

Difference of NO flux between two chamber techniques (dynamic chamber flux – static chamber flux) versus time of day. Vertical lines indicate one standard deviation of the NO flux measurements made from both chamber techniques.

Conclusions The dynamic flow-through chamber system has been successfully developed in response to a need to measure emissions of nitrogen, sulphur, and carbon compounds for a variety of field applications. Moreover, similar chamber systems have also been deployed to measure uptake of nitrogen, sulphur, ozone, and hydrogen peroxide gases by crops and vegetation to examine

108

V. P. Aneja et al

atmospheric-biospheric interactions. Experimental results have been effectively validated by independent flux models. Acknowledgements Funding of this research project was provided by the U.S. Department of Agriculture (USDA) as a part of the National Research Initiative (NRI) under Contract No. 2003-05360. Mention of trade names or commercial products does not constitute endorsement or recommendation for use. References Anderson, I.C. and J.S. Levine, Simultaneous field measurements of biogenic emissions of nitric oxide and nitrous oxide, J. Geophys. Res., 92, 965-976, 1987. Aneja, V.P., Dynamic Studies of Ammonia Uptake by selected Plant Species Under Flow Reactor conditions, Ph.D. Thesis, North Carolina State University, Raleigh, N.C., pp. 216, 1976. Aneja, V.P., G.C. Murray, and J. Southerland, Atmospheric nitrogen compounds: Emissions, transport, deposition, and assessment, EM, April 22-25, 1998. Aneja, V.P., J.P. Chauhan, and J.T. Walker, Characterization of atmospheric ammonia emissions from swine waste storage and treatment lagoons, J. Geophys. Res., 105, No. D9, 11535-11545, 2000. Aneja, V.P., B. Bunton, J.T. Walker, and B.P. Malik, Measurement and analysis of atmospheric ammonia emissions from anaerobic lagoons, Atmos. Environ., 35, 1949-1958, 2001a. Aneja, V.P., P.A. Roelle, G.C. Murray, J. Southerland, J.W. Erisman, D. Fowler, W.A.H. Asman, and N. Patni, Atmopheric nitrogen compounds II: emissions, transport, transformation, deposition, and assessment. Atmos. Environ., 35, 1903-1911, 2001b. Aneja, V.P., P.A. Roelle, and Y. Li, Effect of environmental variables on NO emissions from agricultural soils, Phyton (Austria), 4, 29-40, 2001c. Aneja, V.P., B.P. Malik, Q. Tong, D. Kang, and J.H. Overton, Measurement and Modelling of ammonia emissions at waste treatment lagoon-atmospheric interface, Water, Air, and Soil Pollution, 1, 177-188, 2001d. Arkinson, H.L., Measurements, Modeling, and Analysis of Ammonia Flux from Hog Waste treatment technologies, M.S. Thesis, pp.129, 2003. Claiborn, C.S., Atmospheric-Biospheric Interactions of Ambient H2O2: Climatology at Mt. Mitchell, N.C., and Transport into Needles of Red Spruce, Ph.D. Thesis, North Carolina State University, 1991. Claiborn, C. S., and V.P. Aneja. Transport and fate of reactive trace gases in red Spruce needles, part 1: Uptake of gaseous hydrogen peroxide as measured in controlled chamber flux experiments. Environmental Science and Technology 27, 2585-2592, 1993. Danckwerts, P.V., Gas-Liquid Reactions, McGraw-Hill Book Company, New York, 1970. Gao, W., M.L. Wesely, D.R. Cook, and T.J. Martian, Eddy-correlation measurements of NO, NO2 and O3 fluxes, in Proceedings of Measurement of Toxic and Related Air Pollutants, pp. 146-150, Air and Waste Manage. Assoc., Pittsburgh, PA, 1996. Hagerman, L.M., V.P. Aneja, and W.A. Lonneman, Characterization of non-methan hydrocarbons in the rural southeast United States, Atmos. Environ., 31, 4017-4038, 1997.

Dynamic Chamber System to Measure Gaseous Compounds Emissions

109

Heck, W.W.,O.C. Taylor, and D.T. Tingey, Assessment of Crop Loss from Air Pollutants, Elsevier Applied Science, London, U.K. p. 552, 1987. Hill, F.B, V.P. Aneja, and R.M. Felder, Atechnique for measurement of biogenic sulfur emission fluxes, J. Environ. Sci. Health, A13(3), 199-225, 1978. Kim, D.S., V.P. Aneja, and W.P. Robarge, Characterization of nitrogen oxide fluxes from soil of a fallow field in the coastal piedmont of North Carolina, Atmos. Environ., 28, 1129-1137, 1994. Levenspiel, O., Chemical Reaction Engineering, John Wiley &Sons, New York, 1962. Li, Y., V.P. Aneja, S.P. Arya, J. Rickman, J. Brittig, P. Roelle, and D.S. Kim, Nitric oxide emissions from intensively managed agricultural soil in North Carolina, J. Geophys. Res., 104, 26,115-26,123, 1999. Roelle, P.A. and V.P. Aneja, Characterization of ammonia emissions from soils in the upper coastal plain, North Carolina, Atmos. Environ., 36, 1087-1097, 2002. Roelle, P.A., V.P. Aneja, B. Gay, C. Geron, and T. Pierce, Biogenic nitric oxide emissions from cropland soils, Atmos. Environ., 35, 115-124, 2001. Roelle, P.A., V.P. Aneja, J. O’Connor, W. Robarge, D.S. Kim, and J.S. Levine, Measurement of nitrogen oxide emissions from an agricultural soil with a dynamic chamber system, J. Geophys. Res., 104, 1609-1619, 1999. Rogers, H.H., H.E. Jeffries, and A.M. Witherspoon. Measuring air pollutant uptake by plants: nitrogen dioxide. Journal of Environmental Quality 8, 551-557, 1979. Tabachow, R.M., P.A. Roelle, J.J. Pierce, and V.P. Aneja, Soil nitric oxide emissions: lab and field measurements and comparison, Environmental Engineering Science, 19(4), 205-213, 2002. U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park, NC, November 1994, EPA-453/R-94-080A.

Chamber Studies on the Photolysis of Aldehydes John C. Wenger Department of Chemistry and Environmental Research Institute, University College Cork, Ireland Key Words: Photolysis, Aldehydes, Photooxidation

Introduction Aldehydes are emitted directly into the atmosphere from a variety of natural and anthropogenic sources and are also formed in situ from the atmospheric degradation of volatile organic compounds (VOCs). The atmospheric fate of aldehydes is controlled by photolysis and reaction with hydroxyl (OH) or nitrate (NO3) radicals and, in the case of unsaturated compounds, reaction with ozone (Atkinson, 1994). The photolysis of aldehydes is of particular importance because it is a source of free radicals in the troposphere, and thus may significantly influence the oxidizing capacity of the lower atmosphere (Finlayson-Pitts and Pitts, 1986). Aldehydes absorb in the near ultraviolet range and typically exhibit a weak absorption band in the wavelength range 240-360 nm as a result of a symmetry forbidden n-S* transition. The minimum wavelength of sunlight that reaches the troposphere is 290 nm, which is of sufficient energy to cause photolysis of aldehydes. The importance of photolysis relative to other loss processes such as reaction with OH, NO3 or O3 can be determined by comparison of the rates of all the competing degradation pathways. If photolysis is found to be important then the products need to be identified and mechanisms constructed in order to provide a more complete picture of the photolysis process. In this paper an overview of the application of simulation chamber studies to the photolysis of aldehydes is presented and a brief review of the kinetic and mechanistic data currently available is also provided. Finally, results from some of the first studies of the photolysis of aromatic aldehydes in an outdoor simulation chamber are presented and discussed. Determination of Photolysis Rate Coefficients The photolysis of a species X follows first order kinetics; j X + hQ o Products

and is described by the photolysis rate coefficient j. The rate of removal of X by photolysis is thus; -d[X]/dt = j[X] The photolysis rate coefficient (in s-1) depends on the absorption cross-section (ı, base e in cm2 molecule-1), the quantum yield for photolysis (ij, in molecule photons-1) and the solar actinic flux (F, in photons cm-2 s-1), all of which are wavelength dependent; j = ı(Ȝ) ij(Ȝ) F(Ȝ) 111 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 111–119. © 2006 Springer. Printed in the Netherlands.

(,)

J. C. Wenger

112

Under atmospheric conditions where a range of wavelengths at Ȝ > 290 nm are present then the atmospheric photolysis rate coefficient is best described by the following expression; Ȝi

j

¦ ı(Ȝ)M(Ȝ) F (O )

(II)

Ȝ 290 nm

where ı(Ȝ) and ij(Ȝ) are averaged over the wavelength interval ǻȜ and F is summed over the same wavelength range. The sum is carried out from 290 nm to some wavelength Ȝi at which either the quantum yield or absorption cross-section becomes negligible (Finlayson-Pitts and Pitts, 1986). Calculation of j requires the values of ı and ij to be known. While the experimental determination of absorption cross-sections is fairly straightforward, the measurement of quantum yields is not, mainly due to the rapid secondary reactions of radicals. In fact, there is no quantum yield data available for the vast majority of aldehydes. As a result, calculations of photolysis rates are often carried out in which it is assumed that ij(Ȝ) = 1.0. This, of course, only produces a value for the maximum photolysis rate. In addition, because the actinic flux depends on many factors, including geographical location, time and season, photolysis rates are normally calculated for certain conditions; e.g. at noon on a cloudless day at a location 40°N on the Earth’s surface on July 1. The problems associated with calculating photolysis rates can be overcome by the experimental determination of j values in outdoor simulation chambers, such as the European Photoreactor (EUPHORE), in Valencia, Spain (Becker, 1996). The decay of an aldehyde when irradiated by natural sunlight can be measured directly by FTIR spectroscopy or gas chromatography and the j values determined from a simple first order kinetic plot in the form of equation (III); ln[X]0/[X]t = -jt

(,II)

To ensure that photolysis is the only loss process for the aldehyde experiments can be carried out in the presence of an excess concentration of a radical scavenger such as cyclohexane. In cases where the high concentration of a scavenger is undesirable, e.g. because it causes saturation in the infrared absorption spectrum, a tracer compound, such as di-n butyl ether, can be used to correct for the measured decay of the aldehyde in order to obtain the j value. Typical starting concentrations used in photolysis experiments at EUPHORE are [aldehyde] = 0.5-1.5 ppmv, [scavenger] = 10-50 ppmv or [tracer] = 0.1-02. ppmv (Wenger et al., 2004 and Magneron et al., 2002). The experimentally determined value of j is, of course, only measured under one particular set of atmospheric conditions. Since the photolysis rate depends on the intensity and spectral distribution of the sunlight it is useful to quote the photolysis rate coefficient relative to a conventional measure of solar light intensity, such as j(NO2). The value for j(aldehyde)/j(NO2) can be used to calculate the photolysis rate of the aldehyde under a range of atmospheric conditions. Information on the quantum yield for photolysis of the aldehyde by sunlight can de obtained by comparison of the experimentally determined value of j with the maximum theoretical value. The latter is calculated from solar flux intensity measurements during the experiment, known absorption cross-section data and assuming ij(Ȝ) = 1.0. The ratio of the

Chamber Studies on the Photolysis of Aldehydes

113

measured j value to the calculated value gives the effective quantum yield, ijeff, over the atmospheric absorption range of the aldehyde. During the last 5 years, the photolysis of over twenty aldehydes has been performed at EUPHORE. Many of these studies were performed as part of the EU-funded research project RADICAL (Moortgat, 2000). A summary of the reported j values and effective quantum yields is provided in Table 1 along with the calculated lifetimes due to photolysis (IJphot = 1/jexp) and reaction with OH radicals (IJOH = 1/kOH[OH]). Table 1.

Summary of photolysis rate coefficients (jexp) measured at EUPHORE. IJphot (hr)

IJOHa (hr)

(2.9 ± 2.7) × 0.06 ± 0.05 10-6

96

17

(1.1 ± 0.1) × 10-5

3.8 × 10-5

0.28 ± 0.04

29

14

Butyraldehyde

(1.0 ± 0.2) × 10-5

5.1 × 10-5

0.20 ± 0.04

29

12

i-Butyraldehyde

(3.7 ± 0.1) × 10-5

5.2 × 10-5

0.71 ± 0.02

7.5

11

n-Pentanal

(1.6 ± 0.2) × 10-5

5.4 × 10-5

0.30 ± 0.02

17

10

-5

-5

0.72 ± 0.03

7.3

12

3-Methylbutyraldehyde

(1.25 ± 0.1) × 10

-5

0.27 ± 0.01

22

12

Pivaldehyde

(1.45 ± 0.1) × 10-5 2.6 × 10-5

0.56 ± 0.05

19

10

n-Hexanal

(1.65 ± 0.3) × 10-5 5.9 × 10-5

0.28 ± 0.05

17

15

n-Nonanal

(1.15 ± 0.2) × 10-5 4.9 × 10-5

0.23 ± 0.03

24

10

Glyoxal

(1.05 ± 0.3) × 10-4 2.7 × 10-3

0.04 ± 0.01

2.6

25

Glycolaldehyde

(1.15 ± 0.3) × 10-5 8.6 × 10-6

Compound

jexp (s-1)

jmax (ij = 1) (s-1)

Acetaldehyde

(2.9 ± 2.7) × 10-6

Propionaldehyde

2-Methylbutyraldehyde

Methacrolein

(3.8 ± 0.1) × 10

-5

-6

< × 10

-6

5.2 × 10

4.7 × 10

ijeff

1.32 ± 0.30

24

28

5.2 × 10

-4

< 0.004

> 144

8

-4

< 0.004

> 144

13 7.4

Acrolein

< × 10

4.3 × 10

trans-Crotonaldehyde

(1.2 ± 0.2) × 10-5

4.0 × 10-4

0.03 ± 0.01

24

Pinonaldehyde

(1.15 ± 0.1) × 10-5 8.0 × 10-5

0.14 ± 0.01

24

3.1 6

a) OH rate coefficients taken from (Atkinson, 1994), lifetime calculated assuming [OH] = 1 × 10 molecule cm-3.

The straight chain C3-C9 aldehydes all have similar photolysis rates and values for ijeff in the range 0.20-0.30. For these compounds the atmospheric lifetimes for photolysis by sunlight are longer than the lifetimes for reaction with OH radicals. In contrast, the Įbranched aldehydes possess significantly higher values for ijeff and their photolysis lifetimes are shorter than those for reaction with OH radicals. Interestingly, the ijeff values for the

J. C. Wenger

114

unsaturated aldehydes, methacrolein, acrolein and trans-crotonaldehyde are very small, even though these compounds possess absorption spectra extending towards the visible. In many atmospheric chemistry models quantum yields are assumed to be unity, which, as shown in Table 1, is seldom the case. Thus it appears that the importance of photolysis in the degradation of the aldehydes and their contribution to radical formation in the atmosphere is often over-estimated. However, details concerning the production of radicals from the atmospheric photolysis of aldehydes are best obtained from mechanistic studies. Mechanisms for the Photolysis of Aldehydes There are a number of theoretical reaction pathways in the photolysis of aldehydes, as demonstrated below for n-hexanal (Tadiü et al., 2001); CH3CH2CH2CH2CH2CHO + hȞ ĺ ĺ ĺ ĺ ĺ ĺ

n-C5H11 + HCO (Norrish Type I) n-C5H12 + CO C4H8 + CH2=CHOH (Norrish Type II) C3H7 + CH2CH2CHO C5H10 + HCHO CH3CH2CH2CH2CH2CO + H

(1) (2) (3) (4) (5) (6)

Reactions (1) and (3) are believed to be the most important under atmospheric conditions. The Norrish Type I process results in fragmentation into free radicals, whilst the Norrish Type II process, which is common to molecules with a Ȗ-hydrogen atom, is an intramolecular rearrangement that results in no radical formation. Experiments performed at EUPHORE and in indoor photoreactors as part of the RADICAL project found that the Norris Type I process is dominant for n-butanal and smaller straight chain aldehydes, but the Norrish Type II process is the major pathway for the photolysis of n-pentanal and n-hexanal. Simulation Chamber Studies on the Photolysis of Aromatic Aldehydes Aromatic aldehydes are emitted into the atmosphere as primary pollutants from automobile exhausts and are also formed in situ from the OH radical initiated oxidation of alkyl aromatic compounds. Tolualdehydes and dimethylbenzaldehydes are produced from the gas-phase reaction of OH with xylenes and trimethylbenzenes respectively (Calvert et al., 2002). The subsequent atmospheric fate of these aromatic aldehydes is controlled by reaction with OH and NO3 radicals and photolysis by sunlight. Kinetic studies of the gas-phase reactions of the tolualdehydes and dimethlybenzaldehydes have been recently performed (Thiault et al., 2002 and Clifford, 2004) and preliminary studies of the photolysis of the tolualdehydes have also been reported (Volkamer et al., 2000 and Thiault et al., 2001). In this work the photolysis of the 3 tolualdehydes and 6 dimethylbenzaldehydes (DMBAs) were investigated during July 2003. The measured photolysis rate coefficients have been used to calculate tropospheric lifetimes and provide an assessment of the relative importance of photolysis as an atmospheric loss process for the compounds. Experiments were performed in Chamber B of the EUPHORE facility. Reactants were introduced into the chamber by gently heating a known amount of the compound in a glass impinger and flushing the vapour into the chamber via a stream of purified air. Concentrations ranging from 131 to 511 ppbv of reactant were used during different experiments. The

Chamber Studies on the Photolysis of Aldehydes

115

reactant was allowed to mix for approximately one hour and its rate of deposition to the walls of the chamber was measured during this period. Photolysis was then initiated by opening the protective housing and thereby exposing the chamber to sunlight. The solar actinic flux was measured over the range 290-520 nm with a spectral resolution of 1 nm FWHM using a calibrated spectroradiometer (Bentham TM300). A full spectral scan took 420-430 seconds. The duration of photolysis was typically 2-3 hours. The temperature inside the chamber increased slightly as the experiments progressed but was always within the range 296-308 K. Chemical analysis was performed throughout the experiments by in situ FTIR spectroscopy using an optical path length of 553.5 m. Infrared spectra in the range 600-4000 cm-1 were obtained using a Nicolet Magna 550 FTIR spectrometer operated at a resolution of 1 cm-1 using a broad band MCT detector and derived from the co-addition of 270 scans collected over 5 minutes. Additional chemical detection was provided by a gas chromatograph (Fisons 8160) equipped with flame ionisation and photoionisation detectors (FID and PID). The chromatograph was operated using a 30 m DB-624 fused silica capillary column (J&W Scientific, 0.32 mm i.d., 1.8 Pm film). Air was sampled from the chamber into a 3 ml sampling loop and then injected onto the column. The temperature of the column was held isothermal (120º C for tolualdehydes and 150ºC for dimethylbenzaldehydes). The reactants were quantified using calibrated reference infrared spectra and gas chromatographic sensitivity factors obtained by introducing known volumes of pure materials into the chamber. The leak rate from the chamber was determined daily by adding about 20 ppbv of the unreactive tracer gas SF6 and measuring its loss by FTIR spectroscopy. The derived correction factors were applied to determine the amounts of reactants consumed and products formed. Table 2.

Experimental details for the photolysis of 2,4-DMBA at EUPHORE. Date

9th July 2003

Initial Concentration (ppbv) Irradiation time j (NO2)average (s-1) Results k(2,4-DMBA)FTIR (s-1) k(2,6-DMBA)GC-PID (s-1) kSF6 (s-1) kwall (s-1) j(2,4-DMBA)FTIR (s-1) j(2,4-DMBA)GC-PID (s-1) j(2,4-DMBA)average (s-1) j(2,4-DMBA)/j(NO2))

303 2hr 39min (7.47 ± 0.75) × 10-3 (2.14 ± 0.02) × 10-4 (2.43 ± 0.04) × 10-4 (9.19 ± 0.88) × 10-6 (2.44 ± 0.76) × 10-5 (1.90 ± 0.03) × 10-4 (2.19 ± 0.03) × 10-4 (2.05 ± 0.03) × 10-4 (2.74 ± 0.27) × 10-2

Except for j(NO2), quoted errors are twice the standard deviation arising from the least squares fit of the data. For j(NO2) the estimated error is 10%.

A total of eight experiments were performed at EUPHORE in July 2003. The experiments were conducted for at least two hours during the middle of the day. The loss of the aromatic aldehydes was monitored using FTIR spectroscopy and GC-PID. The measured decay rates of the compounds were corrected for wall loss and dilution. Concentration-time

J. C. Wenger

116

profiles and kinetic plots were generated for each of experiments. Data for the photolysis of one of the compounds, 2,4-dimethylbenzaldehyde (2,4-DMBA), are shown in Figures 1 and 2, with the associated experimental details listed in Table 2. Similar data were obtained for all the aromatic aldehydes. A summary of the photolysis rate coefficients obtained in the EUPHORE experiments is given in Table 3. The values quoted for j(aldehyde) are the average of the data obtained by FTIR spectroscopy and GC-PID in several experiments. The ratio j(aldehyde)/j(NO2) was also determined and can be used to calculate the photolysis rate of the aromatic aldehydes under different light conditions in the real atmosphere. The effective quantum yields obtained for o-, m- and p-tolualdehyde were determined by comparing the measured j values with the maximum theoretical values. The latter were calculated from solar flux intensity measurements during the experiments, known absorption cross-section data (Thiault et al., 2004) and assuming ij(Ȝ) = 1.0. There is no gas-phase absorption cross-section data available for the dimethylbenzaldehydes. However, inspection of the solution phase spectra indicates that the light absorption characteristics will be similar to those of the tolualdehydes. From the measured experimental photolysis rates of 2,3-, 2,4-, 2,5- and 2,6- DMBA it is expected that the effective quantum yields will be similar to that of o-tolualdehyde and will be close to unity. For 3,4- and 3,5DMBA, which did not photolyse at EUPHORE, it is expected that the quantum yields will be similar to m- and p-tolualdehyde, i.e. about 0.03.

350

0.008

Chamber Closed

Chamber Open

0.007

250 0.006 200

0.005

150

0.004

100

FTIR GC-PID j(NO2)

0.003 0.002

50

0.001

0 08:24

j(NO2) (s-1)

2,4-DMBA (ppbv)

300

0.009

0.000 08:52

09:21

09:50

10:19

10:48

11:16

11:45

12:14

Time of day (hh:mm)

Figure 1.

Concentration-time profile and j(NO2) during the photolysis of 2,4-DMBA at EUPHORE on 9th July 2003.

The following compounds were readily photolysed by sunlight; o-tolualdehyde, 2,3-, 2,4-, 2,5- and 2,6-DMBA. The photolysis rate coefficients are all remarkably similar, with values for j(aldehyde) in the range 2.0 – 2.7 × 10-4 s-1. For the compounds that did not undergo photolysis (m- and p-tolualdehyde, 3,4- and 3,5- DMBA), the rate of decay of the compounds is equivalent to the wall loss or dilution rate. The j values determined for these

Chamber Studies on the Photolysis of Aldehydes

117

compounds are therefore upper limits. The photolysis rate coefficients obtained for o-, m- and p-tolualdehyde in this study are in good agreement with those previously reported (Volkamer et al., 2000 and Thiault et al., 2001). As this is the first study on the photolysis of the DMBAs, no comparative data exists. 0

ln[2,4-DMBA]0/[2,4-DMBA]t - kwallt

-0.2 -0.4 -0.6 -0.8 -1 -1.2 -1.4 -1.6

FTIR GC-PID

-1.8 -2 0

1000

2000

3000

4000

5000

6000

7000

8000

9000

Photolysis time (s)

Figure 2.

Photolytic loss rate of 2,4-DMBA at EUPHORE on 9th July 2003.

Table 3.

Results obtained for the photolysis of the tolulaldehydes and DMBAs at EUPHORE.

Compound

o-tolualdehyde m-tolualdehyde p-tolualdehyde 2,3-DMBA 2,4-DMBA 2,5-DMBA 2,6-DMBA 3,4-DMBA 3,5-DMBA

j(aldehyde) (s-1)

j(aldehyde)/j(NO2) -4

(2.15 ± 0.06) × 10 (7.43 ± 0.60) × 10-6 (7.43 ± 0.60) × 10-6 (2.28 ± 0.20) × 10-4 (2.05 ± 0.03) × 10-4 (2.36 ± 0.09) × 10-4 (2.31 ± 0.06) × 10-4 (9.36 ± 0.20) × 10-6 (8.27 ± 0.18) × 10-6

-2

(2.66 ± 0.27) × 10 (0.09 ± 0.01) × 10-2 (0.09 ± 0.01) × 10-2 (3.02 ± 0.30) × 10-2 (2.74 ± 0.27) × 10-2 (3.52 ± 0.35) × 10-2 (3.01 ± 0.30) × 10-2 (0.11 ± 0.01) × 10-2 (0.10 ± 0.01) × 10-2

ijeff

1.19 ± 0.04 < 0.05 < 0.03 1.0* 1.0* 1.0* 1.0* 0.03* 0.03*

* Estimated as no gas-phase data available The photolysis rate coefficients obtained in this work can be used to calculate the tropospheric lifetimes of the aromatic aldehydes with respect to photolysis. The lifetimes listed in Table 4 were calculated from the values of j(aldehyde)/j(NO2) obtained at

J. C. Wenger

118

EUPHORE and assuming a value of j(NO2) = 1 × 10-2 s-1, which is typical of the maximum photolysis rate at solar noon during summer months. The atmospheric lifetimes due to reaction with OH and NO3 are also shown for comparative purposes. For o-tolualdehyde, 2,3-, 2,4-, 2,5- and 2,6-DMBA, photolysis is clearly the dominant atmospheric loss process, whilst for the remaining aldehydes, gas-phase reaction with OH radicals is the most important degradation pathway. Lifetimes (in hours) of the aromatic aldehydes with respect to the various atmospheric loss processes.

Table 4.

Compound

WOH

WNO3

Wphotolysis

o-tolualdehyde m-tolualdehyde p-tolualdehyde 2,3-DMBA 2,4-DMBA 2,5-DMBA 2,6-DMBA 3,4-DMBA 3,5-DMBA

8.7 8.4 8.5 6.8 6.4 6.3 5.8 7.1 6.2

56.7 58.7 58.7

1.0 30.9 30.9 0.9 1.0 0.8 0.9 25.3 27.8

The OH and NO3 lifetimes were calculated from rate coefficients determined by (Clifford, 2004). No rate coefficients are available for the reaction of NO3 with the DMBAs.

The fact that photolysis only occurs for compounds in which the -CHO group is adjacent (ortho) to a methyl group suggests that some form of intramolecular interaction between these functional groups is a significant component of the photolysis mechanism. This aspect is currently being investigated further and photolysis mechanisms are in development. Acknowledgements This author would like to thank the following co-workers; G.M. Clifford, A. Mellouki, G. Le Bras, A. Munoz, M. Martín-Reviejo and K. Wirtz. This work was supported by the European Commission, through the IALSI project (Contract EVR1-CT-2001-40013), the Irish Higher Education Authority (HEA) and the French Programme National de Chimie Atmosphérique (PNCA). The CEAM Foundation is supported by the Generalitat Valenciana and BANCAIXA. References Atkinson, R; Gas-phase tropospheric chemistry of organic compounds, J. Phys. Chem. Ref. Data, Monograph 2 (1994) 1-216. Becker, K.H. (Ed.); The European Photoreactor EUPHORE, Final Report of the EC Project EV5V-CT92-0059, Wuppertal (1996). Clifford, G.M.; Atmospheric chemistry of aromatic aldehydes, PhD thesis, University College Cork (2004). Finlayson-Pitts, B.J., J.N. Pitts Jr.; Atmospheric Chemistry, John Wiley, New York (1986) Magneron, I., R. Thévenet, A. Mellouki, G. Le Bras, G.K. Moortgat and K. Wirtz; A study of the photolysis and OH-initiated oxidation of acrolein and trans-crotonaldehyde, J. Phys. Chem. A 106 (2002) 2526.

Chamber Studies on the Photolysis of Aldehydes

119

Moortgat, G.K. (Ed.); Evaluation of radical sources in atmospheric chemistry through chamber and laboratory studies, Final Report of the EC Project ENV4-CT97-0419, Mainz (2000). Tadiü, J., I. Juraniü and G.K. Moortgat; Photoxidation of n-hexanal in air, Molecules, 6 (2001) 287-299. Thiault, G., A. Mellouki, G. Le Bras and K. Wirtz; The photolysis of aromatic aldehydes in The European Photoreactor, EUPHORE, 4th report, (2001), 29. Thiault, G., A. Mellouki , G. Le Bras; Kinetics of gas phase reactions of OH and Cl with aromatic aldehydes; Phys. Chem. Chem. Phys, 4 (2002) 2194 - 2199. Thiault, G., A. Mellouki , G. Le Bras,A. Chakir, N. Sokolowski-Gomez, D. Daumont; UV-absorption cross sections of benzaldehyde, ortho-, meta-, and para-tolualdehyde, J. Photochem. Photobiol. A: Chem. 162 (2004) 273. Volkamer, R., U. Platt and K. Wirtz; OH reaction rate constants and photolysis frequencies of a series of aromatic aldehydes and phenols in The European Photoreactor, EUPHORE, 3rd report, (2000) 1. Wenger, J.C., S. Le Calvé, H.W. Sidebottom, K. Wirtz, M. Martín-Reviejo and J.A. Franklin; Photolysis of chloral under atmospheric conditions, Environ. Sci. Technol. 38 (2004) 831-837.

Determination of Photolysis Frequencies for Selected Carbonyl Compounds in the EUPHORE Chamber Romeo-Iulian Olariu1, Marius Duncianu1, Cecilia Arsene1, and Klaus Wirtz2 1

2

“Al.I. Cuza” University of Iasi, Faculty of Chemistry, Analytical Chemistry Department, 700506 Iasi, Romania. Centro de Estudios Ambientales del Mediterráneo, Parque Tecnológico, C.\ Charles R. Darwin 14, 46980 Paterna, Valencia, Spain Key Words: Photolysis, Actinic flux, Spectroradiometer, Carbonyl compounds

Introduction Small carbonyl compounds are formed during the photochemical oxidation of many volatile organic compounds (VOC’s), in urban as well as in rural areas. Photolysis and reaction with the OH radical are the most important initiation reactions for the atmospheric degradation of these compounds, and lead to the formation of peroxy radicals in the former case and either stable molecules and/or free radicals in the latter case (Finlayson-Pitts and Pitts, 1999). About 580 actinic flux spectra recorded under different meteorological conditions (clear sky, partially cloudy or overcast) in the EUPHORE smog chambers, have been used to calculate the photolysis frequencies for various small carbonyl compounds that are considered to be important from the atmospheric chemistry point of view. The results are presented here. Experimental a) Actinic Flux measurement (F) The actinic flux spectra were recorded under realistic atmospheric conditions in the large scale outdoor EUropean PHOtoREactor (EUPHORE). The EUPHORE chamber is made from PTFE-Teflon foil. It is half spherical in shape with a volume of about 204 m3 and is mounted on aluminium floor panels. The PTFE-Teflon foil is highly transparent even to short wavelength radiation, with transmissions ranging from 85% in the range 500 - 320 nm to > 75% at 290 nm. A detailed description of the experimental set-up is available elsewhere (Becker, 1996). The light intensity was measured using a calibrated spectroradiometer (Bentham DM300). Both direct and reflected beams of the incident light are optically coupled and pass simultaneously but geometrically separated through a double monochromator. The entire set of actinic flux spectra have been recorded in the range of 290 nm to 520 nm. The spectra were recorded every 5 min with a spectral resolution of 1 nm in the range from 290 nm to 360 nm and 5 nm between 360 to 520 nm. The actinic flux spectra were recorded during an entire year, i.e. under different meteorological conditions, clear sky, scattered clouds or overcast etc. b) Photolysis frequencies calculation (Jx) In practice, rather than integrating over the wavelength spectrum to obtain the total rate of photolysis, the sum of product V(O) × I(O) × F(O) over discrete wavelength intervals 'Oҏwas used. The choice of the width of these intervals was usually based on the intervals for which F(O) was measured. Thus, the total photolysis rate was calculated using the equation:

121 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 121–128. © 2006 Springer. Printed in the Netherlands.

122

R.-I. Olariu et al.

J

Oi

¦O

290

V (O ) u I O u F (O)

(E1)

where V (O ) and I (O ) are the values averaged over a wavelength interval 'O centred at O, and F(O) is actinic flux in photons cm-2 s-1 summed over the wavelength interval'Ocentred at O The numerical sum over the relevant wavelength range in eq. E1 has been performed using a home-produced MS-DOS program. The program contains a data base for the computation of 43 photo-dissociation reactions of 40 different trace gases of atmospheric relevance. For the present study the measured EUPHORE radiation spectra together with published molecular data (Moortgat, 2000) were used for the calculation of the photolysis frequencies of 17 carbonyl compounds.

c ) Calculation of solar zenith angle (T) The solar zenith angle can be calculated in the following manner for any particular location (i.e. latitude and longitude), day of the year and time of the day (Finlayson-Pitts and Pitts, 1999). There are two steps that need to be considered. First, the “local hour angle” (th), which is defined as the angle between the meridian of the observer and that of the sun, in radians, has to be calculated:

th = S[(GMT/12)-1 + (longitude/180)] + EQT

(E2)

where GMT is Greenwich mean time converted from the local time, longitude (in degree, west of the Greenwich meridian are negative) and EQT is the “equation time”, given by: -5

-3

EQT = 7.5 ×10 + 1.868 ×10 cos N -2

–3.2077 ×10 sin N -2

–1.4615 ×10 sin 2N - 4.0849 ×10 -2sin 3N

(E3)

where N (in radian) is defined as:

N= 2Sdn/365

(E4)

The day of the year, dn, is defined as the day number (0 - 364), with 0 corresponding to January 1 and 364 to December 31. The second parameter that is needed for the calculation of the solar zenith angle at a particular time and place is the solar declination, G, defined as the angle between the direction of the sun and the equatorial plain of the earth. The value of G (in radian) can be calculated as follows: -3

G = 6.918 ×10 -0.399912cos N + 0.070257 sin N – -3

-4

- 6.758 ×10 cos 2N + 9.07 ×10 sin 2N -3

-3

–2.697 ×10 cos 3N + 1.480 ×10 sin 3N

(E5)

Determination of Photolysis Frequencies for Selected Carbonyl Compounds

123

with these two parameters the solar zenith angle for any particular time and place can be easily calculated: cos T = sinG ×sin(latitude) + cos G×cos(latitude) ×cos th

(E6)

where Gand th are calculated as already described and latitudes north of the equator (in radian) are positive and south are negative. If all of the input parameters are in radians, T is also obtained in radians and can be converted to degree using 1 rad = 57.2960. For example, at Valencia, Spain, (longitude: 0.50 W, latitude: 39.50 N) on September 1 at noon local time, GMT = 11.0, dn= 243, N = 4.183, EQT = - 0.0016, th = -0.2721 rad, G = 0.149 rad, and cosT = 0.829, giving a solar zenith 0 angle of 0.592 rad, or 33.92 Consequently, the actinic flux data recorded in the EUPHORE chamber during one year will provide useful and complementary data over the range of the zenith angle values pertinent to this local area. Results and discussion The actinic flux spectra were used to calculate the spectral photolysis frequency distributions taking into account the molecular parameters (V(O) x I(O)), which after wavelength integration yield the respective J values. Figure 1 shows typical actinic flux profiles recorded in the EUPHORE chamber at various times of the year. The spectral data have been used to obtain atmospheric photo-dissociation frequencies for several carbonyl compounds. For all the investigated compounds (see Table 1) plots of the calculated photolysis frequencies versus solar zenith angle show a cloud-type distribution of data points. Figure 2 shows an example for acetaldehyde. The strong scattering and lack of a clear linear dependency has probably several causes one of which is likely the variation in the degree of cloudiness over the EUPHORE chamber. Since the photolysis frequency is directly coupled with the solar zenith angle one major objective of the present study was the realization of a relative dependency between the calculated photolysis frequencies and solar zenith angle. To obtain a clear dependency a statistical treatment of all the calculated photolysis frequencies, derived from all the actinic spectra recorded in the EUPHORE chamber has been performed. Figure 3 shows an example of the complete dataset for acetaldehyde. The statistical treatment has allowed a clear dependency between the calculated photolysis frequency and the solar zenith angle to be established. The result obtained for acetaldehyde statistical treatment after is presented graphically in Figure 4. The error bars represent the statistical (1V) error only. Photolysis frequencies have been calculated in the range of 19º to 71.5º solar zenith angles for 17 carbonyl compounds. The calculated photolysis frequencies obtained for the different zenith angles as derived from all the EUPHORE actinic flux spectra measurements are presented in Table 1. Using the data obtained in the present work a simple analytical expression has been derived for Jx(T) which captures more than 75% of all the measurements up to a 71.50 solar zenith angle with an error band of 14%. The general analytical expression obtained is as follows: Jx(T) = [m u exp(-n u sec(T)] s-1 2

E7

The empirical parameters m, n and r obtained from fits of the calculated photolysis frequencies as a function of solar zenith angle are listed in Table 2.

124

R.-I. Olariu et al.

Figure 1.

Actinic flux profiles recorded in the EUPHORE chamber for a 400 zenith angle at the following dates and times: a) on 24.03.2003 at 13:56 h; b) on 20.06.2002 at 16:00 h; c) on 20.03.2001 at 13:09 h; d) 20.03.2002 at 12:55 h.

Figure 2.

Plot of the calculated photolysis frequencies of acetaldehyde from the entire data set against solar zenith angle.

Determination of Photolysis Frequencies for Selected Carbonyl Compounds

Figure 3.

125

Calculated photolysis frequency profile for one year derived from data recorded using the EUPHORE outdoor chamber.

Photolysis frequenciey of acetaldehyde (s-1)

4.0e-5

3.0e-5

2.0e-5

1.0e-5

0.0 0

Figure 4.

10

20

30 40 50 60 Solar zenith angle (degree)

70

80

90

Plot of the calculated photolysis frequencies of acetaldehyde (obtained after statistical treatment) versus solar zenith angle.

R.-I. Olariu et al.

126 Table 1.

-1

Calculated photolysis frequencies for carbonyl compounds (in s ) at different zenith angles derived from the EUPHORE chamber measurements.

Compound

exp

acetaldehyde CH3CH2CHO n-Butyraldehyde i-Butyraldehyde 3-Methybutyraldehyde Pentanal Pinonalddehyde Pivalaldehyde Glyoxal Glycolaldehyde Acrolein Methacrolein Crotonaldehyde Caronaldehyde 2-Petanone MVK Pyruvic_acid

10 10-5 -5 10 -5 10 -5 10 10-5 10-5 -5 10 -3 10 -6 10 -4 10 -4 10 -4 10 10-5 10-5 -4 10 -4 10

-5

Table 2.

19.10 3.35±0.51 3.76 ±0.58 4.93±0.75 4.98±0.76 4.37±0.68 5.06±0.77 6.70±1.05 2.76±0.43 2.82±0.86 8.27±0.13 4.65±0.73 4.59±0.70 4.40±0.68 3.53±0.57 1.28±0.20 4.55±0.70 2.86±0.45

zenith angle (degree) 32.03 51.83 2.96±0.41 2.09±0.24 3.32±0.46 2.34±0.27 4.31±0.61 3.10±0.36 4.35±0.62 3.12±0.36 3.90±0.57 2.78±0.35 4.44±0.63 3.20±037 5.86±0.83 4.05±0.50 2.47±0.36 1.74±0.22 2.97±0.37 2.58±0.22 7.17±0.10 4.87±0.65 4.48±0.60 3.54±0.30 4.38±0.59 3.42±0.30 4.20±0.56 3.29±0.28 3.07±0.46 2.07±0.28 1.10±0.16 0.74±0.10 4.35±0.58 3.41±0.30 2.79±0.38 2.24±0.19

71.53 0.87±0.14 0.97±0.01 1.31±0.22 1.31±0.22 1.21±0.20 1.37±023 1.62±0.26 7.41±0.12 1.45±0.39 1.87±0.31 1.82±0.32 1.73±0.30 1.67±0.29 7.88±0.12 0.28±0.04 1.73±0.30 1.17±0.21

Empirical parameters m, n and r2 from E7.

2

compound

m

n

r

acetaldehyde CH3CH2CHO n-Butyraldehyde i-Butyraldehyde 3-Methybutyraldehyde Pentanal Pinonalddehyde Pivaldehyde Glyoxal Glycolaldehyde Acrolein Methacrolein Crotonaldehyde Caronaldehyde 2-Petanone MVK Pyruvic_acid

10-4 -4 10 -4 10 -4 10 -4 10 -4 10 -4 2 x 10 -4 10 -4 9 x 10 -4 10 -4 8 x 10 -4 8 x 10 -4 7 x 10 -4 10 -4 10 -4 8 x 10 -4 5 x 10

0.7291 0.760 0.737 0.741 0.758 0.736 0.807 0.779 0.356 0.869 0.484 0.509 0.503 0.889 0.884 0.503 0.457

0.96 0.98 0.98 0.98 0.97 0.98 0.98 0.97 0.98 0.98 0.98 0.98 0.98 0.97 0.98 0.99 0.99

Determination of Photolysis Frequencies for Selected Carbonyl Compounds

127

In Table 3 the mean of the all calculated photolysis frequencies for the various compounds obtained from the analysis of the actinic flux measurements performed using EUPHORE chamber facilities is compared with results from other similar studies. In the estimates a quantum yield of unity has been assumed. The errors are a combination of the statistical errors and the errors associated with uncertainties in the molecular parameters. The effective quantum yields, Ieff, in Table 3 were determined by division of the measured individual photolysis frequencies (J3) by the photolysis frequencies calculated in this study (J1) assuming a quantum yield of unity. For some of the carbonyl compounds the determined Ieff is significantly below unity. The photolytic lifetimes obtained using the present photolysis frequencies are also presented in Table 3. Table 3.

compound

Photolysis frequencies and effective quantum yield Ieff for some carbonyl compounds as derived from the EUPHORE facilities and found in the literature. exp -5

J1(X), s

-1

estimated this work

Acetaldehyde 10 2.26±0.84 -5 CH3CH2CHO 10 2.53±0.95 -5 3.34±1.23 n-Butyraldehyde 10 -5 3.37±1.25 i-Butyraldehyde 10 -5 2.97±1.19 3-Methyl-butyraldehyde 10 -5 3.44±1.27 Pentanal 10 -5 4.43±1.75 Pinonaldehyde 10 -5 1.86±0.71 Pivaldehyde 10 -3 2.54±0.47 Glyoxal 10 -6 8.52±1.58 Glycolaldehyde 10 -4 3.62±0.97 Acrolein 10 -4 Methacrolein 10 3.51±0.98 -4 Crotonaldehyde 10 3.37±0.93 -6 2-Pentanone 10 8.28±3.46 -4 MVK 10 3.49±0.97 -4 Pyruvic acid 10 2.26±0.58 a b Moortgat, 2000, Wirtz, 1999.

J2(X), s

-1

estimated literature a

4.9 a 3.8 a 5.1 a 5.2 a 4.7 a 5.4 a 8.0 a 2.6 a 2.7 a 8.6 a 4.3 a 5.2 a 4.0 b 8.6 a 5.2 a 2.3

J3(X), s

-1

Ieff

measured values a

0.29±0.27 a 1.10±0.10 a 1.00±0.20 a 3.70±0.10 a 1.25±0.10 a 1.60±0.20 a 1.15±0.10 a 1.45±0.10 a 0.10±0.03 a 11.4±2.5 a <0.01 a <0.01 a 0.11±0.01 b 0.6 a <0.01 a 1.00±0.15

0.12 0.46 0.29 1.09 0.42 0.46 0.26 0.78 0.039 1.33 <0.003 <0.003 0.03 0.07 <0.003 0.44

Remarks In the present work a large set of actinic spectra recorded in the EUPHORE chamber under various atmospheric conditions has been obtained and used for the calculation of photolysis frequencies of 17 organic carbonyl compounds. From a statistical analysis of the photolysis frequencies calculated for the compounds an analytical form for JX(T) has been derived. For unsaturated compounds (methyl vinyl ketone, methacrolein, acrolein and crotonaldehyde) Ieff is negligible, although those cxompounds possess absorption spectra reaching the near visible.

128

R.-I. Olariu et al.

In atmospheric models currently used, quantum yields are often assumed to be unity, which is seldom the case, as is shown in this study. This leads to overestimation of the calculated photodissociation rates and the associated radicals formed in the photolysis processes. These new data will reduce the present uncertainties associated with photolysis processes, so that future assessments can be made with more confidence. Acknowledgments Financial supports of this work from the European Commission within the 5th Framework Programme project IALSI and action travel funding from Joint Research Centre, Ispra, Italy, to NATO EST. ARW 980164 Research Workshop are gratefully acknowledged. References Finlayson-Pitts, B. J. and J. N. Pitts. Jr.; Chemistry and Physics of the Upper and Lower Atmosphere, Wiley, New York, 1999. Becker, K. H.; The European Photoreactor “EUPHORE”, Final report , EC contract EV5V-CT92-0059, 1996. Moortgat, G. K.; Evaluation of Radical Sources in Atmospheric Chemistry trough Chamber and Laboratory Studies, “RADICAL”, Final Report on EU Project, EC contract ENV4-CT97-0419, 2000. Wirtz K.; lecture presented at the combined US/Germany Environmental Chamber Workshop, Riverside, 1999.

Remote Sensing of Glyoxal by Differential Optical Absorption Spectroscopy (DOAS): Advancements in Simulation Chamber and Field Experiments R. Volkamer1, I. Barnes2, U. Platt3, L. T. Molina1, and M. J. Molina1 1

Department of Earth, Atmosphere and Planetary Sciences, MIT, 77, Massachusetts Av. 54-1320, Cambridge, MA 02139, USA 2 Bergische Universität Wuppertal, FB C - Physikalische Chemie, Gauss Strasse 20, 42119 Wuppertal, Germany 3 Institut für Umweltphysik, University of Heidelberg, D-69120 Heidelberg, Germany Key Words: Glyoxal, DOAS, Atmosphere, Sources, Photolysis, Isoprene

Introduction Air pollution in many large cities is linked with the photochemical transformation of primary pollutants like VOCs (volatile organic compounds) and NOx, which in the presence of sunlight foster the formation of secondary pollutants including ozone (O3) and secondary organic aerosol (SOA) (Finlayson-Pitts and Pitts 2000; Molina and Molina 2002). ‘Photochemical smog’ has adverse effects on human health (Kunzli et al. 2000; Evans et al. 2002), the ecosystem (Middleton et al. 1950; Gregg et al. 2003) and regional climate (Lelieveld et al. 2001; Ramanathan and Crutzen 2003). While the crucial role of the atmospheric chemistry of VOCs in the formation of urban smog is well established, the performance of traditional indicators for VOC chemistry like O3, Ox (sum of O3 and nitrogen dioxide (NO2)), formaldehyde (HCHO), and methylvinylketone (MVK) is affected by direct vehicle emissions, which contribute significantly to the atmospheric concentrations of these species; in the case of O3, high concentrations of nitric oxide (NO) efficiently scavenge this gaseous pollutant during morning rush hour (Finlayson-Pitts and Pitts 2000). Hence, these trace gases may be ambiguous indicators for VOC chemistry in urban air, where the simultaneous presence of multiple pollutant sources complicates the development of robust control strategies to reduce O3 and aerosol levels. Recently, glyoxal (CHOCHO) was detected for the first time directly in the atmosphere by means of differential optical absorption spectroscopy (DOAS) (Volkamer et al. 2005a). These time-resolved glyoxal measurements revealed atmospheric glyoxal concentrations to be essentially unaffected by vehicle emissions in Mexico City, and demonstrated that glyoxal is a unique indicator of fast VOC photochemical oxidation in urban air. Glyoxal is the smallest Į-dicarbonyl hydrocarbon and a mutagenic oxidation product (Kasai et al. 1998) from the oxidation of numerous VOCs (Calvert et al. 2000; Calvert et al. 2002). In a recent study using DOAS at the European Photo Reactor (EUPHORE) the glyoxal yield from the OH-radical initiated oxidation of benzene, toluene and p-xylene have been investigated (Volkamer et al. 2001). These time-resolved glyoxal yields allowed an improvement

129 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 129–141. © 2006 Springer. Printed in the Netherlands.

130

R. Volkamer et al.

in the chemical codes of aromatic VOC oxidation (Bloss et al. 2004b). However, quantitative mechanistic conclusions were limited by the then uncertain DOAS calibration, namely the UV/vis absorption cross-section of glyoxal. An improved UV/vis absorption cross-section of glyoxal has since been recorded and now allows this limitation to be overcome (Volkamer et al. 2005b) and has enabled further improvements in mechanistic conclusions about aromatic oxidation. During the day, rapid photolysis of glyoxal is a source of radicals. Glyoxal loss due to photolysis and reaction with OH-radicals limit its atmospheric lifetime during the day (Plum et al. 1983; Atkinson 2000). The atmospheric lifetime with respect to these gas-phase glyoxal sinks was recently quantified experimentally in Mexico City and found to be considerably shorter than previously believed, i.e. about 1-2 hours during most of the day (Volkamer et al. 2005a). On ‘cloudless’ days, glyoxal photolysis accounts for about two thirds of the gas-phase glyoxal loss; OH-radical reactions are relatively more important in the presence of clouds, at elevated solar zenith angles (SZA>80°), and at night. The SZA dependence of glyoxal photolysis is discussed in more detail in this contribution. In this contribution the re-evaluated yields from the OH-radical initiated oxidation of benzene, toluene, p-xylene, and initial results of new simulation chamber experiments on ‘prompt’ glyoxal formation from isoprene oxidation are presented. A detailed discussion of sources, sinks and their uncertainties to model atmospheric concentrations of glyoxal is presented, and exemplifies how basic research in environmental simulation chambers besides giving input for photochemical models also triggers advancements with measurement techniques for field observations. The integration of laboratory and field observations by models in turn will guide future research on atmospheric chemical processes. Experimental During April 2003, a field campaign was held to investigate the formation of photochemical smog in the Mexico City Metropolitan Area (MCMA). Long-Path DOAS (LPDOAS) was used to detect glyoxal using its unique specific narrow-band (< 5 nm) absorption structures in the visible spectral range, that enable DOAS (Platt 1994) to separate trace gas absorptions from broadband molecule and aerosol extinction in the open atmosphere. As part of the MCMA-2003 field campaign, two LP-DOAS instruments were deployed at the CENICA super-site. Only data from DOAS2 are presented here, which faced an array of retroreflectors at Museo Fuego Nuevo, Cerro de la Estrella. The total atmospheric path length was 4420m at a mean height of 70m above ground level. Atmospheric spectra were recorded at a spectral resolution of 0.4 nm Full Width at Half Maximum (FWHM) about every 2 to 15 minutes. Glyoxal was retrieved in the spectral range between 420 and 465 nm, covering three prominent glyoxal absorption lines using a 5th order polynomial high pass filter and non-linear least squares fitting of glyoxal, NO2, O4, H2O, and two lamp-reference spectra using the Windoas evaluation software (Fayt and van Roozendael 2001). As reference spectra available high-resolution absorption cross-section spectra (see (Alicke et al. 2003) and references therein) were degraded to match the instrument resolution. The DOAS detection of glyoxal is inherently calibrated from

Remote Sensing of Glyoxal by Differential Optical Absorption Spectroscopy

131

knowledge of its differential absorption cross-section (Volkamer et al. 2005b), which for the three covered absorptions lines were 5.5, 2.5, and 1.7×10-19 cm2 at 455, 440, and 428 nm, respectively. Absorption features with differential optical densities of 8×10-4 could be detected, corresponding to a mean detection limit of 3×109 molec/cm3 or 0.15 ppbv (1 ppbv = 1 parts per billion by volume = 1000 pptv § 1.92 ×1010 molec/cm3 in Mexico City). The formation of glyoxal in the early stage of the OH-radical initiated oxidation of isoprene (‘prompt’ glyoxal) in the presence of NOx was also investigated using DOAS at the EUPHORE chamber. A general description of EUPHORE is available in the literature (Klotz et al. 1998). A detailed description of the DOAS instrument at EUPHORE and the relative yield approach used here is available in Volkamer et al. (2001) and Volkamer (2001). In brief, measured product concentrations are corrected for losses through reaction with OH-radicals, photolysis and wall-deposition. The ratio of the yield of a target product (ɎTAR) relative to the known yield of a reference product (ɎREF) is determined from: ɎTAR / ɎREF = [TAR]corr / [REF] corr From a plot of the corrected concentration of the target product as a function of the corrected concentration of the reference product a straight line is expected if both products are formed as primary products. Any secondary formation of the product leads to deviations from the linear relationship. The yield ratio is determined from the slope of the data. The relative yield method is employed in this work to determine the primary glyoxal yield relative to methacrolein, which is an established primary product from the OH-radical initiated oxidation of isoprene. For the discussion of glyoxal sources from VOC oxidation, glyoxal yields were calculated for VOCs where no measurements are reported in the literature. Suitable single VOC compound subsets of the Master Chemical Mechanism MCMv3.1 (Saunders et al. 2003; Bloss et al. 2004a; Bloss et al. 2004b) were prepared. The code was initialised using the following initial conditions: [VOC]i=50ppbv, SO2=1ppbv, CO=120ppbv, CH4=1800ppbv, O3=50ppbv, H2O2=2ppbv, NO2=7ppbv, variable NO (1-100 ppbv) and mid-latitude conditions, and run for three consecutive days. Glyoxal yields were derived as the overall moles of glyoxal formed (loss corrected) divided by the moles of VOCi reacted after one day. Results and Discussion As part of this MCMA-2003 campaign, glyoxal was detected for the first time directly in the atmosphere (Volkamer et al. 2005a). As is visible from the concentration time profile shown in Figure 1, the rapid variation of glyoxal was well resolved by these time resolved DOAS measurements, which present useful data to validate the performance of photochemical models that are used as tools for the design of control strategies to improve air quality in Mexico City.

132

Figure 1.

R. Volkamer et al.

Concentration time profile of glyoxal measured on 10 April 2003 in Mexico City. Letters indicate times of the corresponding atmospheric spectrum shown on the right. Spectra have been vertically shifted for clarity; thick lines indicate scaled reference spectra of glyoxal; error bars indicate 2-sigma level. Reproduced from (Volkamer et al. 2005a).

Previous indirect measurements of glyoxal by cartridge collection/chromatography methods do not resolve the rapidly changing temporal variation of glyoxal in urban air. These are highly specialized and time intensive techniques (Grosjean et al. 1999; Moortgat et al. 2002), which involve pre-concentration of glyoxal and many other carbonyls on a solid sorbent (cartridge) covered with either DNPH (2,4-dinitrophenylhydrazine) or PFBHA (O-(2,3,4,5,6pentafluorobenzyl) hydroxylamine) derivatisation agent, followed by solvent extraction and gas/liquid chromatography separation and mass spectrometric/UV-visible spectroscopic detection. Possible complications may arise if no differentiation between hydroxyacetaldehyde and glyoxal can be made (Grosjean et al. 1996a). The time resolution with such measurements depends on sample accumulation time (one to four hours) and detection limits depend on sample volume and typically range between 50 - 200 ppt. The detection of glyoxal by DOAS in the open atmosphere poses several advantages, including the absence of sampling artefacts, selective detection, inherent calibration, and the combination of high time-resolution (few minutes) with reasonable sensitivity (on the order of 100 ppt). Atmospheric glyoxal concentrations measured by DOAS and other methods are listed in Table 1. Glyoxal levels in the MCMA are well within the concentration range observed in other

Remote Sensing of Glyoxal by Differential Optical Absorption Spectroscopy

133

urban environments, and indeed do not seem exceptionally high, suggesting that routine DOAS measurements of glyoxal are feasible also in other locations. Table 1.

Atmospheric concentrations of glyoxal.

Reference

Location

Airmass

(Volkamer et al. 2005a) (Ho and Yu 2002) (Kawamura et al. 2000) (Grosjean et al. 1996b)

Mexico City, Mexico Hong Kong, China Los Angeles, USA Los Angeles mean Long beach Central LA Azusa Claremont Las Vegas, USA Summer Winter Porto Allegre, Brazil Giesta, Portugal Pabsthum, Germany Georgia, USA

polluted urban polluted urban polluted urban polluted urban

(Jing et al. 2001)

(Grosjean et al. 1999) (Borrego et al. 2000) (Moortgat et al. 2002) (Lee et al. 1995)

Mixing ratio [ppb] < 0.15 – 1.82 0.5 – 4.1 0.04 – 0.95 0.78 ± 0.85 <0.12 – 0.8 <0.12 – 2.2 <0.12 – 3.0 0.4 – 3.8

urban

urban semi-urban Semi-rural rural

0.12 – 0.42 0.09 – 0.21 0.3 0.52 – 2.42 0.01 – 0.12 0.02 – 0.19

A compilation of glyoxal sources from VOC oxidation is presented in Table 2. Experimental data on the glyoxal branching ratio is given where available; average values are listed if several values are available (Calvert et al. 2000; Calvert et al. 2002). For aromatic VOC elevated concentrations of NO and NO2 were recently demonstrated to affect the relative importance of the various oxidation routes (Klotz et al. 2002; Volkamer et al. 2002; Atkinson and Arey 2003). This influence is also visible in the glyoxal yields (see e.g. Table 5.8 of Volkamer (2001)), and values in Table 2 reflect yields measured under close to atmospheric NOx concentrations. For VOCs marked with an asterix, average yields of glyoxal were calculated using the MCMv3.1 code as described above. As is apparent from Table 2, aromatic VOC show the highest glyoxal yields. The uncertainty of the re-evaluated yields (Volkamer et al. 2001; Volkamer et al. 2005b) is reduced by 50% and no longer limits mechanistic conclusions. Assuming a glyoxal/methylglyoxal ratio of 3:2 from toluene (Smith et al. 1998), and a nitrate yield from the bicycloalkylperoxi radical + NO reaction of 10% (Jenkin et al. 2003; Saunders et al. 2003), this glyoxal yield corresponds to a branching ratio of the bicycloalkyl radical pathway from benzene and toluene of 32% and 57.5%. Including quantified ring-retaining product yields (Klotz et al. 1998; Smith et al. 1998; Volkamer et al. 2002) about 85±8% and 82±6% of the primary oxidation routes are identified (molar basis, errors indicate 2-sigma uncertainty level). This is significantly below unity, indicating that a currently unassigned reaction pathway is needed in order to close the carbon balance in the early stages of the oxidation of benzene and toluene.

134 Table 2.

R. Volkamer et al. Sources of glyoxal from the oxidation of VOCs.

Glyoxal precursor VOC

Oxidant

Glyoxal yield

Reference

benzene OH 32 ± 5 (Volkamer et al. 2005b) benzene* OH 36 estimate this work benzene** OH 53 - 70 estimate this work toluene OH 30.6 ± 6 (Volkamer et al. 2005b) p-xylene OH 31.9 ± 5 (Volkamer et al. 2005b) m-xylene OH 7.9 ± 0.4 (Bandow and Washida 1985a) o-xylene OH 8±4 (Bandow and Washida 1985a) 1,2,4-trimethylbenzene OH 7.8 ± 0.5 (Bandow and Washida 1985b) 1,2,3-trimethylbenzene OH 7.2 ± 0.1 (Bandow and Washida 1985b) ethylbenzene** OH 52 – 58 estimate this work n-propylbenzene** OH 54 – 56 estimate this work i-propylbenzene** OH 55 – 60 estimate this work o-ethyltoluene** OH 14 – 18 estimate this work m-ethyltoluene** OH 2.5 – 4.7 estimate this work p-ethyltoluene** OH 39 – 42 estimate this work acetylene* OH 63.5 estimate this work 2-methyl 1,3-butadiene (isoprene) OH 3 this worka 2-methyl 1,3-butadiene (isoprene)** OH/O3 7–8 estimate this workb 2-methyl 1,3-butadiene (isoprene)** OH/O3 ”2 (Paulson and Seinfeld 1992) ȕ-pinene** OH/O3 4.5 – 5.5 estimate this work ethene O3 0.44 (Calvert et al. 2000) ethene** OH/O3 < 1 – 3.6 estimate this work propene O3 8.3 ± 0.6 (Calvert et al. 2000) 1-butene O3 0.3 ± 0.5 (Calvert et al. 2000) cis-2-butene O3 11.3 ± 5.9 (Calvert et al. 2000) trans-2-butene O3 14.7 ± 6.2 (Calvert et al. 2000) 2-methyl-2-butene O3 4.5 ± 2.1 (Calvert et al. 2000) 2-methylpropene O3 0.5 ± 0.4 (Calvert et al. 2000) 1,3-butadiene O3 0.4 ± 0.3 (Calvert et al. 2000) 1-pentene O3 2±2 (Calvert et al. 2000) 2-methyl-1-butene O3 0.4 ± 0.8 (Calvert et al. 2000) 2-methyl-2-butene O3 4.5 ± 2.1 (Calvert et al. 2000) cis-4-octene O3 5.8 ± 0.3 (Calvert et al. 2000) trans-4-octene O3 3 ± 0.2 (Calvert et al. 2000) trans-3,4-dimethyl-3-hexene O3 11.8 ± 1.6 (Calvert et al. 2000) 3,4-diethyl-2-hexene O3 8.5 ± 1.2 (Calvert et al. 2000) Cyclohexene O3 15 ± 1 (Calvert et al. 2000) n-butane** OH < 0.1 estimate this work 2,2-dimethyl butane** OH < 0.2 estimate this work 3-methyl hexane** OH < 0.003 estimate this work 2-methylpentane** OH < 0.5 estimate this work acrolein OH 7.5 ± 2.5 (Magneron et al. 2002) trans-crotonaldehyde OH 16 ± 4 (Magneron et al. 2002) ethyl 1-propenyl ether O3 16.7 ± 0.2 (Grosjean and Grosjean 1999) 4-hexene-3-one O3 6.4 ± 0.3 (Grosjean and Grosjean 1999) cis-3-hexenyl acetate O3 0.3 ± 0.1 (Grosjean and Grosjean 1999) *calculated primary glyoxal yield; **calculated overall glyoxal formation (primary and secondary oxidation product); a experimental value for the primary glyoxal yield from the isoprene + OH reaction; b not including primary glyoxal formation (see a).

Remote Sensing of Glyoxal by Differential Optical Absorption Spectroscopy

135

Despite the significant progress which has been made in recent years on the understanding of aromatic oxidation routes, care must be taken with predictions of glyoxal from chemical codes. Beyond the first generation of oxidation products very little information is as yet available about the identity and fate of the products following ring-cleavage, even for the most abundant alkylbenzenes (Bloss et al. 2004a; Bloss et al. 2004b). Oxidation routes are highly speculative as can be exemplified for benzene, where the primary glyoxal yield is identical to the overall glyoxal formation after several hours of oxidation (Volkamer et al. 2001). This is only partly reproduced by the MCM code, which reflects the primary glyoxal yields but overestimates glyoxal formation in the later stage of the experiment. The predicted overall glyoxal yields from benzene range between 53% and 70%, depending on the initial conditions. Higher yields are predicted at lower initial NOx concentrations. This apparently significant secondary formation of glyoxal from benzene is not in agreement with the experimental observations, as is the apparently reversed NOx dependence of the yields. In analogy to toluene and p-xylene, glyoxal yields from benzene are expected to be lower in the absence of NOx (see e.g. Table 5.8 of Volkamer (2001)). While the formation of glyoxal from aromatic VOC is likely over-predicted by the MCM code (in part due to the lower re-evaluated glyoxal yields), glyoxal formation is under estimated from alkenes like cis- and trans-2-butene, for which glyoxal is not a predicted oxidation product. Glyoxal is formed with high yield from acetylene, which will likely be only a minor source in urban areas, and relatively more important in remote air, when more reactive VOCs are depleted. The glyoxal source from alkanes and oxygenates is expected to be minor as a consequence of low glyoxal yields and/or VOC concentrations.

Figure 2.

Glyoxal and methacrolein formation from the OH-radical initiated oxidation of isoprene during the experiment ISO2 at EUPHORE (27.03.02).

136

R. Volkamer et al.

Glyoxal is formed from the oxidation of biogenic VOC like isoprene and ȕ-pinene in nonnegligible amounts Yields are, however, currently not well established. Existing isoprene oxidation schemes (Paulson and Seinfeld 1992; Saunders et al. 2003; Geiger et al. 2003) rationalize glyoxal formation exclusively via the further oxidation of stable intermediate compounds (as secondary glyoxal). Recent quantum calculations have predicted isoprene to form glyoxal as a primary oxidation product (Dibble, 2002). The proposed reaction mechanism involves rapid E/Z isomerization of excited OH-isoprene adducts, followed by 1,5-H-shift of the E-isomer and subsequent glyoxal formation. The hypothesis of primary glyoxal formation the oxidation of isoprene was tested in an experiment at EUPHORE, where glyoxal and methacrolein were measured at high timeresolution by DOAS. Figure 3 shows the data from the first 50mins of reaction time. Relative to the established primary oxidation product methacrolein, glyoxal concentrations increase in a fixed relationship, indicating that glyoxal is indeed formed as a primary oxidation product in the OH-radical initiated oxidation of isoprene. The proposed novel chemistry would appear to be operative. A value of 0.014 has been determined for the ratio of the primary glyoxal and methacrolein yields. Assuming a methacrolein yield from isoprene of 21%, the primary glyoxal yield is about 3%. This novel chemistry needs be incorporated into isoprene oxidation schemes, and will lead to higher predicted glyoxal concentrations in areas with active isoprene chemistry. In Figure 3 the effect of spectral resolution on the absorption features of glyoxal is shown in panel A. The high-resolution absorption cross-section (Volkamer et al. 2005b) was convoluted to lower resolution, using a Gaussian shape line function of 1nm full width half maximum (see inset of panel A). Glyoxal photolysis as a function of wavelength was calculated as described in (Volkamer et al. 2005b) for solar zenith angles of 0, 60 and 80 degrees. In panels B and C, data for overhead sun are drawn as solid lines (dark shaded area in panel C, SZA = 0 degrees), 60 degrees as dashed lines (light grey shaded area), 80 degrees as dotted lines (meshed area), and integrated photolysis frequencies (J-values) correspond to J0=1.17×10-4, J60=5.33×10-5 and J80=1.22×10-5 s-1. While the J0/J60-ratio matches the ratio of photon actinic fluxes of visible light (near 415nm) within 25-30%, differences become larger at higher SZA, i.e. the J60/J80 ratio deviates from the ratio of visible photon actinic fluxes by more than a factor of ten, reflecting that most glyoxal photolysis takes place at shorter wavelengths. In panel C, the photolysis rate is shown accumulated over wavelength (right scale). About one third to half of glyoxal photolysis takes place at wavelengths below 335 nm, where photolysis quantum yields are relatively well known. Current best estimates of quantum yields (RADICAL 2001) suggest that about half to two thirds of glyoxal photolysis takes place at wavelengths > 335 nm, where quantum yields are considerable more uncertain. Despite recent progress in our understanding of glyoxal photolysis quantum yields, most glyoxal photolysis is taking place at wavelengths where quantum yields are currently not well established. A better understanding of the wavelength (and pressure) dependence of quantum yields is desirable.

Remote Sensing of Glyoxal by Differential Optical Absorption Spectroscopy

Figure 3.

137

Glyoxal absorption cross-section and photolysis frequencies (J-values). (A) High resolution absorption cross-section of glyoxal (dotted line) is convoluted to 1nm resolution (solid line) by integrating with a Gaussian shape line function of about 1nm width (inset). (B) photon actinic flux for overhead sun (SZA = O˚ , solid ဨ line), and SZA = 60˚ (dashed line) and SZA = 80˚ (dotted line) and photolysis quantum yield (grey shaded). (C) Glyoxal photolysis as a function of wavelength for SZA = O˚, 60˚ and 80˚ (shaded areas) and accumulated glyoxal photolysis (lines).

Conclusions Measurements of glyoxal (CHOCHO) by Differential Optical Absorption Spectroscopy (DOAS) are feasible on a routine basis in the Mexico City Metropolitan Area (MCMA) and other

138

R. Volkamer et al.

urban environments. The DOAS measurement of glyoxal combines the advantage of real-time detection in the open atmosphere, absence of sampling losses, high time-resolution, reasonable sensitivity, and inherent calibration from the knowledge of the absorption cross-section. Available high-resolution spectra enable the routine detection of glyoxal in most urban air masses and possibly also rural locations. Glyoxal is formed from the oxidation of numerous VOCs. Current modelling capabilities to predict the sources of glyoxal from VOC oxidation require revision of VOC chemical codes. In urban air, where ozone concentrations tend to be low, aromatic VOC chemistry is the dominant glyoxal source and glyoxal formation is likely over estimated as a result of significant, though experimentally not verified, secondary glyoxal from the MCMv3.1 code. In rural areas, the prediction of glyoxal sources is highly uncertain due to the unknown NOx dependence of glyoxal yields. Also, proposed reaction pathways in the isoprene + OH reaction, which lead to the formation of glyoxal as a primary product, have been experimentally verified. While the primary glyoxal yield is low (about 3%), this novel chemistry may be a potentially important glyoxal source in areas of high isoprene emissions. Current isoprene oxidation schemes do not reflect this novel chemistry and likely underestimate glyoxal production. Uncertainties in the sources of glyoxal from VOC chemistry may be hard to detect from model predictions that do not adequately represent the faster than expected glyoxal photolysis. Quantification of gas-phase sink reactions of glyoxal would benefit from a re-determination of the OH-radical rate constant of glyoxal, knowledge of the NO3-radical rate constant as well as a better understanding of the wavelength dependent photolysis quantum yields. Additional glyoxal losses from dry deposition and transfer to aerosols are expected (Jang et al. 2002; Kalberer et al. 2004), but too little information is as yet available to assess the relevance of these sinks for atmospheric residence times. In order to accurately predict glyoxal concentrations in the atmosphere, specific laboratory experiments to improve chemical codes with respect to sources and sinks of glyoxal are needed. Glyoxal measurements present a very useful indicator molecule for volatile organic compound (VOC) chemistry, and provide novel means to test predictions of photochemical models in urban and possibly rural environments. Clearly, the feasibility of routine DOAS measurements of glyoxal opens the possibility to advance our understanding of atmospheric chemical processes in urban and remote environments, and reproducing atmospheric observations presents a challenge for the future. Acknowledgements Financial support from the German BMBF (Program on Atmospheric Research AFO2000), US-National Science Foundation (NSF) and Mexican Metropolitan Environmental Commission (CAM) is gratefully acknowledged. R.V. gratefully acknowledges successive fellowships by the Camille and Henry Dreyfus Foundation and the Alexander von Humboldt Foundation.

Remote Sensing of Glyoxal by Differential Optical Absorption Spectroscopy

139

References Alicke, B., A. Geyer, A. Hofzumahaus, F. Holland, S. Konrad, H.W. Patz, J. Schafer, J. Stutz, A. Volz-Thomas, and U. Platt; OH formation by HONO photolysis during the BERLIOZ experiment, Journal of Geophysical Research-Atmospheres, 108 (2003) 8247-8263. Atkinson, R.; Atmospheric chemistry of VOCs and NOx, Atmospheric Environment, 34 (2000) 2063-2101. Atkinson, R. and J. Arey; Atmospheric Degradation of Volatile Organic Compounds, Chemical Review, 103 (2003) 4605-4638. Bandow, H. and N. Washida; Ring-Cleavage Reactions of Aromatic-Hydrocarbons Studied by Ft-Ir Spectroscopy .2. Photooxidation of Ortho-Xylenes, Meta-Xylenes, and Para-Xylenes in the Nox-Air System, Bulletin of the Chemical Society of Japan, 58 (1985a) 2541-2548. Bandow, H. and N. Washida; Ring-Cleavage Reactions of Aromatic-Hydrocarbons Studied by Ft-Ir Spectroscopy .3. Photooxidation of 1,2,3-Trimethylbenzenes, 1,2,4-Trimethylbenzenes, and 1,3,5-Trimethylbenzenes in the Nox-Air System, Bulletin of the Chemical Society of Japan, 58 (1985b) 2549-2555. Bloss, C., V. Wagner, A. Bonzanini, M.E. Jenkin, K. Wirtz, M. Martin-Reviejo, and M.J. Pilling; Evaluation of detailed aromatic mechanisms (MCMv3 and MCMv3.1) against environmental chamber data, Atmospheric Chemistry and Physics Discussion, 4 (2004a) 5683-5731. Bloss, C., V. Wagner, M.E. Jenkin, R. Volkamer, W.J. Bloss, J.D. Lee, D.E. Heard, K. Wirtz, M. Martin-Reviejo, G. Rea, J.C. Wenger, and M.J. Pilling; Development of a detailed chemical mechanism (MCMv3.1) for the atmosheric oxidation of aromatic hydrocarbons, Atmospheric Chemistry and Physics Discussions, 4 (2004b) 5733-5788. Borrego, C., P. Gomes, N. Barros, and A.I. Miranda; Importance of handling organic atmospheric pollutants for assessing air quality, Journal of Chromatography A, 889 (2000) 271-279. Calvert, J.G., R. Atkinson, K.H. Becker, R.H. Kamens, J.H. Seinfeld, T.J. Wallington, and G. Yarwood (Eds.) (2002), The mechanisms of atmospheric oxidation of aromatic hydrocarbons, Oxford University Press, Oxford. Calvert, J.G., R. Atkinson, J.A. Kerr, S. Madronich, G.K. Moortgat, T.J. Wallington, and G. Yarwood (Eds.) (2000), The Mechanisms of Atmospheric Oxidation of the Alkenes, Oxford University Press, New York. Dibble, T.S.; Isomerization of OH-isoprene adducts and hydroxyalkoxy isoprene radicals, Journal of Physical Chemistry A, 106 (2002) 6643-6650. DOI: 10.1021/jp025682m Evans, J., J. Levy, J. Hammit, C. Santos Burgoa, and M. Castillejos (2002), Health Benefits of Air Pollution Control, in Air Quality in the Mexico Megacity, edited by Molina, M.J. and L.T. Molina, chapter 4, 105-136, Kluwer Academic Publishers, Dordrecht, The Netherlands. Fayt, C. and M. van Roozendael; WinDoas 2.1 – Software User Manual, (2001) Finlayson-Pitts, B. and J.N. Pitts (Eds.) (2000), Chemistry of the Upper and Lower Atmosphere, Academic Press, San Francisco. Geiger, H., I. Barnes, J. Bejan, T. Benter, and M. Spittler; The tropospheric degradation of isoprene: an updated module for the regional atmospheric chemistry mechanism, Atmospheric Environment, 37 (2003) 15031519. Gregg, J.W., C.G. Jones, and T.E. Dawson; Urbanization effects on tree growth in the vicinity of New York City, Nature, 424 (2003) 183-187. Grosjean, E., J.B. DeAndrade, and D. Grosjean; Carbonyl products of the gas-phase reaction of ozone with simple alkenes, Environmental Science & Technology, 30 (1996a) 975-983.

140

R. Volkamer et al.

Grosjean, E., P.G. Green, and D. Grosjean; Liquid chromatography analysis of carbonyl (2,4dinitrophenyl)hydrazones with detection by diode array ultraviolet spectroscopy and by atmospheric pressure negative chemical ionization mass spectrometry, Analytical Chemistry, 71 (1999) 1851-1861. Grosjean, E. and D. Grosjean; The Reaction of Unsaturated Aliphatic Oxygenates with Ozone, Journal of Atmospheric Chemistry, 32 (1999) 205-232. Grosjean, E., D. Grosjean, M.P. Fraser, and G.R. Cass; Air quality model evaluation data for organics .2. C-1-C-14 carbonyls in Los Angeles air, Environmental Science & Technology, 30 (1996b) 2687-2703. Ho, S.S.H. and J.Z. Yu; Feasibility of collection and analysis of airborne carbonyls by on-sorbent derivatization and thermal desorption, Analytical Chemistry, 74 (2002) 1232-1240. Jang, M.S., N.M. Czoschke, S. Lee, and R.M. Kamens; Heterogeneous Atmospheric Aerosol Production by AcidCatalyzed Particle-Phase Reactions, Science, 298 (2002) 814-817. Jenkin, M.E., S.M. Saunders, V. Wagner, and M.J. Pilling; Protocol for the development of the Master Chemical Mechanism, MCM v3 (Part B): tropospheric degradation of aromatic volatile organic compounds, Atmospheric Chemistry and Physics, 3 (2003) 181-193. Jing, L.H., S.M. Steinberg, and B.J. Johnson; Aldehyde and monocyclic aromatic hydrocarbon mixing ratios at an urban site in Las Vegas, Nevada, Journal of the Air & Waste Management Association, 51 (2001) 13591366. Kalberer, M., D. Paulsen, M. Sax, M. Steinbacher, J. Dommen, A.S.H. Prevot, R. Fisseha, E. Weingartner, V. Frankevich, R. Zenobi, and U. Baltensperger; Identification of polymers as major components of atmospheric organic aerosols, Science, 303 (2004) 1659-1662. Kasai, H., N. Iwamoto-Tanaka, and S. Fukada; DNA modifications by the mutagen glyoxal: adduction to G and C, deamination of C and GC and GA cross-linking, Carcinogenesis, 19 (1998) 1459-1465. Kawamura, K., S. Steinberg, and I.R. Kaplan; Homologous series of C-1-C-10 monocarboxylic acids and C-1-C-6 carbonyls in Los Angeles air and motor vehicle exhausts, Atmospheric Environment, 34 (2000) 4175-4191. Klotz, B., S. Sorensen, I. Barnes, K.H. Becker, T. Etzkorn, R. Volkamer, U. Platt, K. Wirtz, and M. Martin-Reviejo; Atmospheric oxidation of toluene in a large-volume outdoor photoreactor: In situ determination of ringretaining product yields, Journal of Physical Chemistry A, 102 (1998) 10289-10299. Klotz, B., R. Volkamer, M.D. Hurley, M.P.S. Andersen, O.J. Nielsen, I. Barnes, T. Imamura, K. Wirtz, K.H. Becker, U. Platt, T.J. Wallington, and N. Washida; OH-initiated oxidation of benzene - Part II. Influence of elevated NOx concentrations, Physical Chemistry Chemical Physics, 4 (2002) 4399-4411. Kunzli, N., R. Kaiser, S. Medina, M. Studnicka, O. Chanel, P. Filliger, M. Herry, F. Horak, V. Puybonnieux-Texier, P. Quenel, J. Schneider, R. Seethaler, J.C. Vergnaud, and H. Sommer; Public-health impact of outdoor and traffic-related air pollution: a European assessment, Lancet, 356 (2000) 795-801. Lee, Y.N., X.L. Zhou, and K. Hallock; Atmospheric carbonyl compounds at a rural southeastern United States site, Journal of Geophysical Research-Atmospheres, 100 (1995) 25933-25944. Lelieveld, J., P.J. Crutzen, V. Ramanathan, M.O. Andreae, C.A.M. Brenninkmeijer, T. Campos, G.R. Cass, R.R. Dickerson, H. Fischer, J.A. de Gouw, A. Hansel, A. Jefferson, D. Kley, A.T.J. de Laat, S. Lal, M.G. Lawrence, J.M. Lobert, O.L. Mayol-Bracero, A.P. Mitra, T. Novakov, S.J. Oltmans, K.A. Prather, T. Reiner, H. Rodhe, H.A. Scheeren, D. Sikka, and J. Williams; The Indian Ocean Experiment: Widespread Air Pollution from South and Southeast Asia, Science, 291 (2001) 1031-1036. Magneron, I., R. Thevenet, A. Mellouki, G. Le Bras, G.K. Moortgat, and K. Wirtz; A Study of the Photolysis and OH-initiated Oxidation of Acrolein and trans-Crotonaldehyde, Journal of Physical Chemistry A, 106 (2002) 2526-2537. Middleton, J.T., J.B. Kendrick, and H.W. Schwalm; Injury to Herbaceous Plants by Smog or Air Pollution, USDA Plant Dis.Rep., 34 (1950) 245-252.

Remote Sensing of Glyoxal by Differential Optical Absorption Spectroscopy

141

Molina, M.J. and L.T. Molina (Eds.) (2002), Air Quality in the Mexico Megacity: An Integrated Assessment, Kluwer Academic Publishers, Dordrecht, The Netherlands. Moortgat, G.K., D. Grossmann, A. Boddenberg, G. Dallmann, A.P. Ligon, W.V. Turner, S. Gab, F. Slemr, W. Wieprecht, K. Acker, M. Kibler, S. Schlomski, and K. Bachmann; Hydrogen peroxide, organic peroxides and higher carbonyl compounds determined during the BERLIOZ campaign, Journal of Atmospheric Chemistry, 42 (2002) 443-463. Paulson, S.E. and J.H. Seinfeld; Development and Evaluation of A Photooxidation Mechanism for Isoprene, Journal of Geophysical Research-Atmospheres, 97 (1992) 20703-20715. Platt, U. (1994), Differential Optical Absorption Spectroscopy, in Monitoring by Spectroscopic Techniques, edited by Sigrist, M.W., chapter 2, 27-84, Wiley & Sons, New York. Plum, C.N., E. Sanhueza, R. Atkinson, W.P.L. Carter, and J.N. Pitts; Oh Radical Rate Constants and Photolysis Rates of Alpha-Dicarbonyls, Environmental Science & Technology, 17 (1983) 479-484. RADICAL; Evaluation of Radical Sources in Atmospheric Chemistry through Chamber and Laboratory studies, ENV4-CT97-0419 (2001) Ramanathan, V. and P.J. Crutzen; New directions: Atmospheric brown "Clouds", Atmospheric Environment, 37 (2003) 4033-4035. Saunders, S.M., M.E. Jenkin, R.G. Derwent, and M.J. Pilling; Protocol for the development of the Master Chemical Mechanism, MCM v3 (Part A): tropospheric degradation of non-aromatic volatile organic compounds, Atmospheric Chemistry and Physics, 3 (2003) 161-180. Smith, D.F., C.D. Mciver, and T.E. Kleindienst; Primary product distribution from the reaction of hydroxyl radicals with toluene at ppb NOX mixing ratios, Journal of Atmospheric Chemistry, 30 (1998) 209-228. Volkamer, R. (2001), A DOAS Study on the Oxidation Mechanism of Aromatic Hydrocarbons under Simulated Atmospheric Conditions, Dissertation, University of Heidelberg, Germany http://archiv.ub.uniheidelberg.de/volltextserver/volltexte/2002/1939/ Volkamer, R., B. Klotz, I. Barnes, T. Imamura, K. Wirtz, N. Washida, K.H. Becker, and U. Platt; OH-initiated oxidation of benzene - Part I. Phenol formation under atmospheric conditions, Physical Chemistry Chemical Physics, 4 (2002) 1598-1610. Volkamer, R., L.T. Molina, M.J. Molina, T. Shirley, and W.H. Brune; DOAS measurement of glyoxal as an indicator for fast VOC chemistry in urban air, Geophysical Research Letters, 32 (2005a) L08806, doi: 10.1029/2005GL022616. Volkamer, R., U. Platt, and K. Wirtz; Primary and Secondary Glyoxal Formation from Aromatics: Experimental Evidence for the Bicycloalkyl-Radical Pathway from Benzene, Toluene, and p-Xylene, Journal of Physical Chemistry A, 105 (2001) 7865-7874. Volkamer, R., P. Spietz, J.P. Burrows, and U. Platt; High-resolution absorption cross-section of Glyoxal in the UV/vis and IR spectral ranges, Journal of Photochemistry and Photobiology A: Chemistry, 172 (2005b) 3542, doi: 10.1016/j.jphotochem.2004.11.011.

Aromatic Hydrocarbon Oxidation: The Contribution of Chamber Oxidation Studies Claire Bloss1, Michael E. Jenkin2, William J. Bloss1, Andrew R. Rickard1, and Michael J. Pilling1 2

1 School of Chemistry, University of Leeds, Leeds LS2 9JT, UK Imperial College London, Silwood Park, Ascot, Berkshire SL5 7PY, UK

Key Words: Aromatic, Oxidation, Mechanism, Troposphere, Ozone, Chamber

Abstract The Master Chemical Mechanism (http://mcm.leeds.ac.uk/MCM) has been updated to version MCMv3.1 in order to incorporate recent kinetic and mechanistic improvements in the understanding of aromatic photo-oxidation. Carefully designed experiments have been carried out in the European Photoreactor (EUPHORE) to investigate key subsets of the toluene system. These results have been used to evaluate the toluene mechanism in MCMv3.1 (and by extension the other aromatic mechanisms in MCMv3.1) and where appropriate to refine the degradation schemes. MCMv3 and MCMv3,1 have also been evaluated using a high quality EUPHORE dataset on the photo-oxidation of benzene, toluene, p-xylene and 1,3,5-trimenthylbenzene. Significant deficiencies have been identified in the mechanisms, in particular: (1) an over-estimation of the ozone concentration, (2) an under estimation of the NO oxidation rate and (3) an under-estimation of OH. The use of MCMv3.1 improves the model-measurement agreement in some areas but significant discrepancies remain. Ideas for additional modifications to the mechanisms and for future experiments to further our knowledge of the details of aromatic photo-oxidation are discussed. Introduction It has been estimated that aromatic compounds, whose major sources are from petrol vehicle exhaust and solvent usage, contribute ~10% to global anthropogenic non-methane hydrocarbon emissions. In regions where there are relatively strong traffic emissions, this percentage can rise substantially (c.f. 21% in the UK (PORG (1997)). The contribution of aromatics to regional photochemical ozone formation is of even greater importance owing to their high reactivity (and hence high photochemical ozone creation potentials (POCPs)) and they can form significant amounts of secondary organic aerosol. For example, in the UK aromatics contribute c.a. 30% to regional ozone production (Derwent et al. (1996)). Therefore, it is important to understand the detailed photochemical degradation mechanisms of aromatic compounds, under a variety of atmospheric conditions, for use in air quality models. The degradation schemes of four aromatic hydrocarbons; benzene, toluene, p-xylene and 1,3,5-trimethylbenzene, have been updated on the basis of new kinetic and mechanistic data from current literature and conference proceedings and are available as part of the latest version of the Master Chemical Mechanism (MCMv3.1) via the MCM website (http://mcm.leeds.ac.uk/MCM). The performance of these schemes concerning ozone formation from tropospheric aromatic oxidation has been evaluated using detailed environmental chamber datasets from the two EU EXACT measurement campaigns at EUPHORE (EXACT I – September 2001 and EXACT II – July 2002 (Pilling et al., 2003)). 143 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 143–154. © 2006 Springer. Printed in the Netherlands.

C. Bloss et al.

144

This comprehensive databank includes photochemical smog chamber studies on the four classes of mono-aromatic compounds above (carried out under suitable atmospheric conditions) as well as a series of experiments looking into specific key areas of aromatic photo-oxidation, focussing on subsets of the toluene system. Results from this system can be transferred to other aromatics as the mechanisms are expected to be similar. Where appropriate, results from this database have been used to refine the mechanisms. Experimental The aromatic validation experiments were performed in chamber A during two campaigns at EUPHORE. A range of techniques was used to measure the concentrations of important species in each system. To ensure high data quality the parent aromatic and ozone were measured with two independent methods, FTIR/GC and UV/FTIR, respectively. NO2 concentrations were measured using DOAS and NO was continuously monitored by a chemiluminescence analyser. LIF was used to measure the concentration of OH radicals during both campaigns, and also measured HO2 during EXACT II (Bloss et al., 2003a). FTIR was used to monitor formaldehyde and nitric acid, GC-ECD was used for PAN and methylglyoxal measurements, DOAS measured glyoxal in addition to NO2, and HPLC detection was used for ring-retained products such as cresols, benzaldehyde and benzoquinones. A comprehensive GC technique from the University of Leeds was used to monitor products such as -dicarbonyls and furanones expected in the ring-opening routes of aromatic oxidation (Hamilton et al., 2003). j-(NO2) was measured using a filter radiometer and this measurement was used to scale all other photolysis rates calculated in the models. Results Development of a detailed chemical mechanism (MCMv3.1) for the atmospheric oxidation of aromatic hydrocarbons The MCM is a near-explicit chemical mechanism developed to describe, in detail, the tropospheric degradation of various important emitted VOCs and is available through the MCM website (http://mcm.leeds.ac.uk/MCM). The first versions of the mechanism used simplified provisional schemes for the poorly known chemistry of aromatic VOC, but in recent years there has been a number of laboratory studies on the atmospheric oxidation of aromatic hydrocarbons and the understanding of the detailed chemistry has improved (e.g. Calvert et al. (2002)). This information has been used to define a rigorous and detailed protocol for the construction of aromatic degradation mechanisms, and these are included in MCM version 3 (Jenkin et al., 2003). However, significant uncertainties remain and development work on these schemes is ongoing. The schemes for benzene, toluene, p-xylene and 1,3,5-trimethylbenzene have been updated following the evaluation experiments performed during EXACT and other data which has become available since the launch of MCMv3. These updated aromatic schemes are now available (via the website) in MCMv3.1. The oxidation pathways of toluene as implemented in the original mechanism (MCMv3) and the updated mechanism (MCMv3.1) are shown in Figure 1. In general, the key areas in which the aromatic mechanisms have been changed are listed below (a more detailed discussion on the updated aromatic chemistry in MCMv3.1 is available in Bloss et al., 2005b):

Aromatic Hydrocarbon Oxidation: The Contribution of Chamber Oxidation Studies

145

x Benzaldehyde yield is slightly lower in the toluene system in MCMv3.1 than MCMv3 in order to better represent the more recent data at low NOx levels (Calvert et al., 2002). x Photolysis rates and subsequent product branching ratios of unsaturated -dicarbonyls (an important source of radicals) have been adjusted according to Thuener et al. (2003) for butenedial and 4-oxo-2-pentenal, and Graedler and Barnes (1997) for 3-hexene-2,5-dione. The photolysis rates of the epoxydicarbonylene products have also been increased as they were originally estimated by analogy with the unsaturated -dicarbonyls. x The breakdown of the carbon skeleton for 2(5H)-furanone (also a photolysis product of butenedial) has been updated and -angelica lactone has been replaced by -angelica lactone to reduce secondary glyoxal formation. These changes were prompted by findings of Volkamer et al. (2001) that secondary glyoxal formation in the toluene system is negligible. x New phenol-type chemistry has been implemented reflecting the lower yield for ringopening in this channel reported by Olariu et al. (2002), and the need for reduced ozone formation found in a cresol photosmog experiment at EUPHORE (e.g. see Figure 5). x Branching ratios for the different oxidation routes of aromatics have been adjusted to reflect the reported yields of glyoxal and of phenol-type compounds at NOx levels appropriate to the atmosphere (Volkamer et al., 2001; Volkamer et al., 2002).

Figure 1.

Schematic representations of toluene oxidation mechanisms. Upper panel MCMv3, lower panel MCMv3.1.

A sensitivity analysis carried out on the MCMv3.1 aromatic chemistry identified the Jdicarbonyls intermediates (co-products of glyoxals in the peroxy-bicyclic ring opening step in MCMv3.1, see Figure 1) as important species affecting OH and O3 concentrations in these systems (Bloss et al., 2005b). 3A comparison of modelled and measured concentrations from a photosmog experiment on butenedial is shown in Figure 2. This experiment and similar

146

C. Bloss et al.

studies on 4-oxo-2-pentenal indicated a number of shortcomings in the mechanisms. NO2 concentration-time profiles are poorly simulated for all experiments, and the differences between experiments under different initial conditions are not well represented. It is clear that the NOxy budget is not well understood. The removal of butenedial is well simulated by MCMv3.1 but the simulated OH and HO2 radical concentrations are significantly lower than the observed values and it seems that a large radical source is missing from the mechanism. This important issue so far remains unresolved. Figure 2 shows a model/measurement comparison from a butenedial photolysis experiment in the absence of NOx. The loss of butenedial is well predicted by MCMv3.1. However, the HO2 concentration is over-estimated by MCMv3.1 by almost an order of magnitude during the early part of the experiment. The time-dependent behaviour is also not well reproduced by the simulation as in the experiment an initial fast increase in concentration is followed by a slower linear increase until the chamber closes, while the simulation shows a fast rise followed by a fall in the HO2 concentration even while the photolysis continues. The photolysis mechanism for butenedial in the absence of NOx as implemented in MCMv3.1 is shown schematically in Figure 4. This indicates that two HO2 radicals should be formed for each molecule of maleic anhydride and glyoxal produced, and while both these product concentrations are over-estimated this is not sufficient to account for the large over-prediction of HO2.

Figure 2. Butenedial photosmog experiment: Figure 3. Butenedial photolysis experiment: comparison of modelled and measured comparison of modelled and measconcentrations of parent compound, ured concentrations of butenedial, HO2, and selected products. O3 , NOx and HOx .

Aromatic Hydrocarbon Oxidation: The Contribution of Chamber Oxidation Studies

Figure 4.

147

Schematic representation of MCMv3.1 butenedial photolysis mechanism in the absence of NOx .

Thuener et al. (2003) propose an alternative butenedial photolysis mechanism based on product experiments. However, this mechanism also requires two HO2 radicals per maleic anhydride molecule formed. One must conclude that a different HO2 formation pathway is operating and/or a significant loss process for HO2 is not included in the mechanism. Concerning the molecular products of butenedial photolysis (Figure 3), the yield of 2(5H)-furanone is well predicted by the simulation, glyoxal and maleic anhydride are overpredicted while the CO yield in the simulation is much lower than observed experimentally. In MCMv3.1 glyoxal and CO are formed as co-products (Figure 4), but this is not consistent with the different yields of these products observed experimentally. Thuener et al. (2003) include a different source of CO in their proposed mechanism, i.e. direct formation from photolysis with a yield of 20%. A possible co-product for direct CO production is acrolein formed by an H-shift and C-C cleavage. The acrolein concentration was below the detection limit of the measurement technique, and its maximum yield was estimated to be ~ 10%. No other direct photolysis products were observed and it was not possible to positively determine the mechanism and co-products for CO formation. In addition to the products for which results are shown in Figure 3, 4-Oxo-but-2-enoic acid and 4-Oxo-but-2-eneperoxoic acid are predicted by MCMv3.1 to be major products in the NOx free conditions of this experiment but were not observed. An unattributed carbonyl absorption was found in the FTIR spectra (Thuener et al., 2003) but the absorption frequency of this unknown does not correspond to that expected for a conjugated carboxylic or peroxoic acid. Although some improvements to our understanding of the photo-oxidation of Jdicarbonyls have been made during the EXACT project the agreement between MCM simulations and chamber experiments, both with and without NOx, remains poor. Particular problems arise concerning the radical yields and seemingly contradictory behaviour is observed. With NOx present, the simulations under-estimate both OH and HO2 while in the absence of NOx the HO2 concentration is over-estimated. In the photosmog systems the NOy budget is not well understood and the simulations are unable to reproduce the different

148

C. Bloss et al.

behaviours seen experimentally at different VOC/NOx ratios and with different J-dicarbonyls. It is therefore difficult to identify mechanistic changes which can simultaneously resolve the discrepancies seen in the two systems (i.e. photosmog and photolysis), based on current general understanding of processes involved in free radical driven VOC oxidation. Nevertheless, overcoming these shortcomings in an area of the mechanism identified as being of key importance in determining ozone and OH concentrations is an important issue for improving our understanding of aromatic chemistry. Another problem encountered in the aromatic photosmog experiments was the difficulty experienced in detecting appropriate amounts of J-dicarbonyls under atmospheric conditions, even using a comprehensive 2D-GC technique (Hamilton et al., 2003). Therefore, improved detection techniques need to be investigated in order to quantify the amounts of J-dicarbonyls formed in these systems. Nevertheless, the EUPHORE chamber dataset on unsaturated Jdicarbonyls provides an important resource for testing future mechanism developments and the work presented here highlights issues that require further investigation. Evaluation of detailed aromatic mechanisms against environmental chamber data As already mentioned, a large, high quality smog chamber dataset on the photooxidation of benzene, toluene, p-xylene and 1,3,5-trimethylbenzene has been assembled from experiments carried out as part of the two EUPHORE EXACT campaigns. The experiments were designed to test sensitive features of detailed aromatic mechanisms, and the dataset has been used to evaluate the performance of the MCMv3 and the updated MCMv3.1. Detailed descriptions of the experimental design and procedure are given in Bloss et al. (2003a) and (2005a). Table 1 summarises the model to measured performance for each experiment in the EXACT photo-smog database. The investigated aromatics form a series of compounds with increasing methyl substitution, and for each compound a range of NOx/VOC ratios are covered by the experiments. The ethene results are included in order to show the level of agreement seen for a relatively well understood system. Figures 5 and 6 show examples of aromatic model-measurement comparisons for benzene and toluene respectively. The O3 peak in the benzene system is greatly reduced in MCMv3.1 compared to MCMv3, and is in good agreement with the measurements, as shown in Figure 5. This substantial improvement is mainly a result of an increased phenol channel yield implemented in MCMv3.1 and an increased ring-retaining product yield in that channel. However, this increase in ring-retaining yield and consequent decrease in radical production leads to a reduction in OH production and oxidative capacity of the system (which was better simulated by MCMv3) as shown in the missing OH calculations in Table 1 and in the decay of benzene in Figure 5. This is indicative of a general problem with the mechanisms; overprediction of the ozone concentration but under-prediction of the reactivity of the system. For the toluene system, model-measurement agreement in terms of O 3 concentration and NO oxidation is improved in MCMv3.1 but the maximum O3 is still over estimated as shown in Figure 6. This mechanism has an increased branching ratio for ring opening in toluene oxidation, and a slightly decreased cresol yield (see Figure 1). These changes, along with the increased photolysis rates of the unsaturated dicarbonyl compounds, tend to increase radical production, particularly early in the experiment, and hence increase O3 generation. However, changes to the substituted phenol (cresol) chemistry, analogous to those discussed above for the benzene system, decrease O3 formation and reduce radical production in the middle of the experiment. The changes balance out in such a way that some decrease in O3

Aromatic Hydrocarbon Oxidation: The Contribution of Chamber Oxidation Studies

149

formation is achieved while increasing the radical production in the early stages of the experiment. However, the amount of “missing” OH needed for the model to reproduce the observed decay of the primary aromatic was calculated over the whole course of the experiment, and in general is higher for the MCMv3.1 simulations indicating an overall reduction in oxidative capacity for this mechanism consistent with the reduced ozone formation potential (see Table 1). Table 1. Measures of model performance concerning the ozone formation and oxidation capacity for MCMv3 and MCMv3.1. For O3 the peak values are taken irrespective of the time they occur, NO oxidation rate is a measure for the timing of the simulation, and the numbers in bold for these measures are the percentage difference between simulated and observed value, relative to the observed value. For missing OH, base value is the amount of new OH produced by the base mechanism over the course of the experiment, “missing” is the amount of extra OH needed to achieve agreement between modelled and observed aromatic decay, and the number in bold is the missing OH as a percentage of the total OH, (base + missing).

In general the updated mechanism, MCMv3.1, shows improved ability to simulate some of the observations from the EXACT EUPHORE datasets and represents our current understanding of aromatic degradation. However, significant discrepancies remain concerning, in particular, ozone formation potential and oxidative capacity of aromatic hydrocarbon systems: x The peak ozone concentration in the benzene experiments is well simulated but for the substituted aromatics the simulations generally over-estimate the peak ozone concentration. x The OH radical production in the mechanisms is too low to account for the OH concentration inferred from the rate of loss of the parent aromatic. The simulations underestimate the oxidative capacity of the aromatic systems. x For the majority of the experiments the simulations under-predict the NO oxidation rate, which is used as a measure of the timing of the simulations. This parameter is closely

150

C. Bloss et al.

related to the processes responsible for the production of ozone and the reactivity of the system.

Figure 5. Model-measurement comparison for Figure 6. Model-measurement comparison for toluene, O3, NO2 and NO in a benzene, O 3 , NO2 and NO in a EUPHORE toluene photosmog EUPHORE benzene photosmog experiment. experiment.

The simultaneous over prediction of O3 concentrations and under prediction of the NO oxidation rate and oxidising capacity of the aromatic systems poses a problem for mechanism development; a reduction in ozone concentration can be achieved by a reduction in peroxy radical concentration, which limits the NO to NO2 conversion, but this will also lead to a reduction in OH production and in the oxidative capacity of the system. A number of possible mechanistic fixes have been investigated and a number of strategies for resolving some of these issues have been proposed (Bloss et al., 2005b): x An increase in OH yield without additional NO to NO2 conversion would improve the simulated concentration-time profiles. A potential OH regeneration channel has been identified, but there is no firm evidence for its occurrence. x Conversion of NO3 to HO2 by a dummy reaction in the model decreases O3 yields but does not significantly affect the aromatic decay rate. x A modelled conversion of NO2 to HONO on the secondary organic aerosol has been shown to improve model-measurement agreement for O3, NOx and toluene decay. For example, Figure 7 shows the effect on the simulations of replacing the usual peroxy radical chemistry by a direct decomposition to ring-opened products and OH. This could occur following a rearrangement of the peroxy radical by a H-shift, as indicated in Figure 8.

Aromatic Hydrocarbon Oxidation: The Contribution of Chamber Oxidation Studies

151

Figure 7. Comparison between experimental Figure 8. Schematic of MCMv3.1 peroxide bicyclic route for toluene oxidation and simulated concentration timeand postulated OH regeneration profiles for MCM3.1 and mechanism pathway. including a speculative OH regeneration mechanism in the peroxidebicyclic route for an EXACT Toluene experiment

It is clear to see that this alteration in the mechanism improves the modelmeasurement agreement. The peak ozone is decreased by 12% compared to MCMv3.1, and the measured peak concentration is over-predicted by 19% by the modified mechanism, 36% by MCMv3.1. At the same time the modelled loss of toluene is faster than in MCMv3.1 as a result of higher OH concentrations, giving a better representation of the observations but still an under-estimate of the reactivity of the system. In the aromatic chamber experiments aerosol particles are formed and chemistry may take place on the surface of these particles. For example, NO2 could partition to the condensed phase and be reduced to HONO by the surface bound species on the secondary organic aerosol formed in the aromatic chamber experiments. Such heterogeneous processes would provide a sink for NOx and a source of radicals (via the photolysis of HONO). A comparison between experimental and simulated concentration time-profiles for MCM3.1 and mechanisms including a conversion of NO2 to HONO on the measured aerosol surface for the toluene experiment shown in Figure 6 is shown in Figure 9. The rate of NO2 reaction on the aerosol was calculated as a function of time, assuming a constant reaction probability of 0.025. The total surface area of the aerosol is also shown and a sharp onset of particle formation was observed just after 11:00. At this point the simulations show a decrease in NO2 concentration and increase in HONO concentration. The OH concentration also increases rapidly, as evidenced by the increase in toluene loss rate. The peak ozone concentration is decreased and under-estimates the measured value. However, this underestimation is a result of the low reactivity in the early part of the simulation before aerosol formation takes place. If an OH source is introduced to reproduce the toluene loss in the first hour of the simulation the NOx and O3 concentrations agree well with the measurements, as shown in Figure 9. However, while the simulations predict a large peak in HONO concentrations no evidence for HONO formation was seen in the FTIR spectra suggesting that the concentration was below approximately 6 ppbv.

152

C. Bloss et al.

Figure 9. Comparison between experimental and simulated concentration time-profiles for MCM3.1, mechanism with a dummy NO3 conversion to HO2 included and mechanisms with artificial OH source added in order to reproduce toluene loss for an EXACT Toluene experiment Summary and Conclusions The detailed oxidation mechanisms of aromatic hydrocarbons in the MCM have been revised and updated based on the latest available experimental data and is available via the MCM website (http://mcm.leeds.ac.uk/MCM) as MCMv3.1. A series of chamber experiments were carried out to investigate the details of key areas of aromatic oxidation mechanisms, and these were primarily focused on toluene oxidation (Bloss et al., 2005b). Comparison of modelled and measured concentrations from photosmog experiments on J-dicarbonyls indicated a number of shortcomings in the mechanisms. NO2 concentrationtime profiles are poorly simulated for all experiments, and the differences between experiments under different initial conditions are not well represented. It is clear that the NOxy budget is not well understood. The simulated HOx concentrations are significantly lower than the observed values and it seems that a large radical source is missing from the mechanism. However, very low HO2 radical concentrations were measured in J-dicarbonyl photolysis experiments in the absence of NOx, these measured concentrations were much lower than HO2 concentrations simulated by MCMv3.1. These important issues remain unresolved. More experiments are needed to determine the yields of reactive oxygenated intermediates formed in aromatic degradation. Ideally these would be carried out with high time resolution as an aid to distinguishing primary and secondary oxidation products. Sensitive, on-line analytical techniques, such as proton-transfer-reaction mass spectrometry, should be employed to detect and quantify such intermediates in chamber experiments carried out under NOx conditions representative of atmospheric levels. Experiments have been carried out in EUPHORE to study the oxidation of benzene, toluene, p-xylene and 1,3,5-trimethylbenzene under conditions appropriate to the atmosphere, and the large, high quality dataset obtained has been used to evaluate the performance of the

Aromatic Hydrocarbon Oxidation: The Contribution of Chamber Oxidation Studies

153

MCMv3 and MCMv3.1 aromatic mechanisms (Bloss et al., 2005a). The updated mechanism, MCM v3.1, shows improved ability to simulate some of the observations from the EUPHORE dataset but significant discrepancies remain, concerning, in particular, ozone formation potential and oxidative capacity of aromatic hydrocarbon systems. The combination of over-prediction of ozone concentration and under-prediction of reactivity poses a problem for mechanism development; a reduction in ozone concentration can be achieved by a reduction in peroxy radical concentration, which limits the NO to NO2 conversion, but this will also lead to a reduction in NO oxidation rate and OH production. A number of possible mechanistic fixes have been investigated (Bloss et al., 2005b) as listed below: x A regeneration mechanism for OH formation without additional NO to NO2 conversion would improve the simulated concentration-time profiles. Flash photolysis experiments with time-resolved OH detection by laser induced fluorescence could be carried out to determine radical yields in the initial stages of aromatic oxidation. The hydroxycyclohexadienyl radical, which in the atmosphere is formed by OH addition to the aromatic ring, could be generated by photolysis of a suitable precursor compound. If the postulated OH regeneration mechanism illustrated in Figure 8 were correct, subsequent reactions with O2 and intramolecular rearrangements would lead to production of OH in the absence of NO. This experiment would require synthesis of a suitable photolytic precursor for the hydroxycyclohexadienyl radical, and careful consideration of the appropriate initial precursor and O2 concentrations. x Conversion of NO3 to HO2 by a dummy reaction in the model decreases O3 yields but does not significantly affect the aromatic decay rate. This indicates that some NO3 chemistry may be missing from the mechanism and could help to explain some of the model-measurement discrepancies. However, any such chemical modifications will not solve the problem of the under-estimation of the oxidative capacity in the system. x A modelled conversion of NO2 to HONO on the secondary organic aerosol has been shown to improve model-measurement agreement for O3, NOx and toluene decay. However, the relatively high HONO concentrations generated have not been observed experimentally and the reactive uptake coefficient required is much higher than the upper limits for such processes suggested by the work of Broske et al. (2003). Further investigation of radical and NOxy budgets under conditions of significant organic aerosol concentration is required to determine the effect of heterogeneous reactions in these aromatic systems. x A sensitivity analysis is planned, to investigate the effect of the uncertainty in the rate of radical generation on ozone formation in the standard photochemical trajectory model (Derwent et al. (1996)). In an environment with many VOC species and radical precursors, this uncertainty may be less significant. The development work discussed here on aromatic degradation schemes has been extended by analogy to update the degradation schemes of twelve other mono-aromatics in the MCMv3.1 with saturated alkyl side chains.

154

C. Bloss et al.

Acknowledgements The authors would like to thank Klaus Wirtz and Montse Martin-Reviejo for their management skills and scientific contributions during the EXACT EUPHORE campaigns and the rest of the EXACT consortium for all their substantial contributions. This work was supported by the European Commission within the EXACT project, Contract No. EVK2-CT1999-00053. References Bloss, C., V. Wagner, A.Bonzanini, M.E. Jenkin, K. Wirtz, M. Martin-Reviejo and M.J. Pilling; Evaluation of the master chemical mechanism (MCM) for aromatic hydrocarbons., EUPHORE Annual Report 2001. (2003a) 34-40. Bloss, C., V. Wagner, A. Bonzanini, M.E. Jenkin, K. Wirtz, M. Martin-Reviejo and M.J. Pilling; Evaluation of detailed aromatic mechanisms (MCMv3 and MCMv3.1) against environmental chamber data., Atmos. Chem. Phys., 5, (2005a) 623-639. Bloss, C., V. Wagner, M.E. Jenkin, R. Volkamer, W.J. Bloss, J.D. Lee, D.E. Heard, K. Wirtz, M. MartinReviejo, G. Rea, J.C. Wenger and M.J. Pilling; Development of a detailed chemical mechanism (MCMv3.1) for the atmospheric oxidation of aromatic hydrocarbons. Atmos Chem. Phys., 5, (2005b) 641-664. Bloss, W.J., J.D.Lee, C. Bloss, K. Wirtz, M. Martin-Reviejo, M. Siese, D.E. Heard and M.J. Pilling; Validation of the calibration of a laser-induced fluorescence instrument for the measurement of OH radicals in the atmosphere. Atmos. Chem. Phys., 4, (2003b) 571-583. Broske, R., J. Kleffmann and P. Wiesen; Heterogeneous conversion of NO2 on secondary organic aerosol surfaces: A possible source of nitrous acid (HONO) in the atmosphere? Atmos. Chem. Phys. 3, (2003) 469474. Calvert, J.G., R. Atkinson, K.H Becker, R.M. Kamens, J.H. Seinfeld, T.J. Wallington and G. Yarwood; The Mechanisms of Atmospheric Oxidation of Aromatic Hydrocarbons, Oxford University Press (2002). Derwent, R.G., M.E. Jenkin and S.M. Saunders; Photochemical ozone creation potentials for a large number of reactive hydrocarbons under European conditions. Atmos. Environ. 30(2), (1996) 181-199. Graedler, F. and I. Barnes; Photolysis of Z-/E-3-Hexenedione. EUPHORE Annual Report 1997, (1997) 148. Hamilton, J.F., A.C. Lewis, C. Bloss, V. Wagner, A.P. Henderson, B.T. Golding, K. Wirtz, K, M. MartinReviejo and M.J. Pilling; Measurements of photo-oxidation products from the reaction of a series of alkylbenzenes with hydroxyl radicals during EXACT using comprehensive gas chromatography. Atmos. Chem. Phys., 3, (2003) 1999-2014. Jenkin, M.E., S.M. Saunders, V. Wagner and M.J. Pilling; Protocol for the development of the Master Chemical Mechanism, MCM v3 (Part B): tropospheric degradation of aromatic volatile organic compounds. Atmos. Chem. Phys. 3, (2003) 181-193. Olariu, R.I., B. Klotz, I. Barnes, K.H. Becker and R. Mocanu; FT-IR study of the ring-retaining products from the reaction of OH radicals with phenol, o-, m-, and p-cresol. Atmos. Environ. 36(22), (2002) 3685-3697. Pilling, M.J. et al.; Effects of the oXidation of Aromatic Compounds in the Troposphere (EXACT, EVK2-CT1999-00053), Final Report., http://www.chem.leeds.ac.uk/exact, (2003). PORG 1997; "Ozone in the United Kingdom", Report prepared by the UK Photochemical Oxidants Review Group. London, HMSO Publications, (1997). Thuener, L.P., G. Rea, and J.C. Wenger; Photolysis of butenedial and 4-oxo-2-pentenal., EUPHORE Annual Report 2001, (2003) 41-46. Volkamer, R., U. Platt, and K. Wirtz; Primary and secondary glyoxal formation from aromatics: Experimental evidence for the bicycloalkyl-radical pathway from benzene, toluene, and p-xylene. J. Phys. Chem. A. 105(33), (2001) 7865-7874. Volkamer, R., B. Klotz, I. Barnes, T. Imamura, K. Wirtz, N. Washida, K.H. Becker and U. Platt; OH-initiated oxidation of benzene - Part I. Phenol formation under atmospheric conditions. Phys. Chem. Chem. Phys. 4(9), (2002) 1598-1610.

FT-IR Kinetic Study on the Gas-Phase Reactions of the OH Radical with a Series of Nitroaromatic Compounds Iustinian Bejan1,2, Ian Barnes1, Romeo Olariu2, Karl Heinz Becker1, and Raluca Mocanu2 1

Bergische Universität Wuppertal, FB-C - Physikalische Chemie, Gaußstraße 20, D-42097 Wuppertal, Germany “Alexandru Ioan Cuza” University of Iasi, Faculty of Chemistry, Department of Analytical Chemistry, Carol I Boulevard 11, 700506 Iasi, Romania

2

Key Words: Rate constant, Gas-Phase, OH Radical, Photolysis, Nitroaromatic, OH formation

Introduction Large quantities of harmful VOCs are emitted into the troposphere from anthropogenic sources (Calvert et al., 2002). Aromatic hydrocarbons are an important class of VOCs present in the atmosphere, which contribute significantly to the chemistry of urban air (Atkinson, 2000; Atkinson and Arey, 2003). Estimations of the global emissions of aromatic hydrocarbon suggest that they comprise between 17-25% of the total anthropogenic NMVOC emissions (Calvert et al., 2002). The degradation of aromatic hydrocarbons is mainly initiated during the day by reaction with the hydroxyl radical (OH). Based on the presently accepted but inadequate mechanism for the atmospheric degradation of aromatic hydrocarbons, it has been calculated that this class of VOC could account for up to 30% of the photooxidant formation in urban areas (Derwent et al., 1996, 1998). Besides the photooxidant formation, this class of hydrocarbon is also assumed to make a significant contribution to secondary organic aerosol (SOA) formation in urban areas (Odum et al., 1996; Forstner et al., 1997; Hurley et al., 2001). This ranks aromatic hydrocarbon as one of the most important classes of hydrocarbons emitted into the urban atmosphere. Formulation of a realistic photo-oxidation mechanism for aromatic hydrocarbons represents one of the most important and challenging problems remaining to be solved in chemical models of tropospheric photooxidant formation. Nitroaromatics such as nitrophenols and nitrocresols have been measured in significant quantities in the atmosphere (Atkinson et al., 1992; Grosjean, 1991), mainly in relation to their potential health effects. Nojima et al. (1975) were the first to detect nitrophenols in the environment and reported the presence of several nitrophenols in the rain water. Nitroaromatics have gained more in atmospheric importance since their phytotoxic properties were discovered (Kawai et al., 1987). Nitroaromatics have been identified in air (Belloli et al., 1999; Tremp et al., 1993), clouds (Lüttke and Levsen, 1997), water (Geissller and Scholer, 1994), soil (Voznakova et al., 1996), fog (Herterich, 1991) and snow (Kawamura and Kaplan, 1986). Nitrated hydroxyaromatics may enter into the atmosphere from both primary and secondary sources. The formation of nitrophenols and nitrocresols in the combustion processes of motor vehicles has been reported by Tremp et al. (1993). Others primary sources may be combustion of coal, wood, manufacture of phenol-formaldehyde resins, pharmaceuticals disinfectants, dyes and explosives (Harrison et al., 2005). Studies in our and other laboratories have shown that an additional important source of these compounds in the atmosphere could be the gas-phase OH-radical initiated photooxidation of aromatic hydrocarbons such as benzene, toluene, phenol, cresols and dihydroxybenzenes in the presence of NOx during the daytime as well as the reaction of NO3 radicals with these aromatics during the night time (Atkinson et al., 1992; Olariu et al., 2002). Once released or 155 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 155–162. © 2006 Springer. Printed in the Netherlands.

I. Bejan et al.

156

produced in the troposphere, these compounds will undergo further oxidation and are also expected to partition between the gas and aerosol phases. Presently very little is known about the atmospheric behaviour of nitrophenols and nitrocresols. The aim of the present study is to increase our knowledge with respect to the fate of this compound class under simulated tropospheric conditions in the laboratory. Recent studies in our laboratory on the photolysis of nitrophenol and nitrocresols with actinic lamps (Philips TL 05/40 W: 320 < Ȝ < 480, Ȝmax = 360 nm) have indicated that these compounds could be important secondary organic aerosol sources (Bejan et al., 2003). In order to assess the importance of this aerosol formation pathway for the atmosphere, kinetic information concerning other atmospheric loss processes, e.g. reactions with OH and NO3 radicals, is required. Experimental The kinetic investigations were carried out in a 1080 L cylindrical quartz-glass photoreactor in synthetic air (296 ± 3 K and 760 ± 10 Torr). As the description given here is limited to the instrumentation actually used in this study, the reader is referred to the literature for a more detailed description of this reaction chamber (Barnes et al., 1994). A White type multiple reflection mirror system, operated at a total optical path length of (484.7 ± 0.8) m, coupled to a Nicolet Nexus FTIR spectrometer equipped with a liquid nitrogen cooled mercury-cadmium-tellurium (MCT) detector, was used for reactant and reference monitoring. IR spectra were recorded at a spectral resolution of 1 cm-1. Between 8 and 12 spectra were recorded in each kinetic experiment, every spectrum comprised of 128 co-added interferograms. Before the kinetic experiment a series of five experiments were performed for every nitrocresol compound in order to obtain the photolysis and wall loss rates. These latter two constants were used in corrections of the OH kinetic rate data. Formation of OH radicals was observed during the photolysis of nitrocresols. Using isoprene as scavenger for OH radicals (kOH = 1.0 × 10-10 cm3 s-1, Atkinson and Arey, 2003) it was possible to calculate the concentration of OH radicals produced by the photolysis and thus calculate a correct rate constant for the photolysis frequency. The photolysis of CH3ONO with 32 superactinic fluorescent lamps (Philips TL05/40 W; 320 nm < O < 480 nm, with Omax=360 nm) in the presence of NO was used as the OH radical source: CH3ONO

+

hQ

CH3O +

CH3O

+

O2

HCHO +

HO2

+

NO

OH

+

NO

HO2 NO2

Gaseous compounds (CH3ONO, NO, ethene, n-butane and isoprene) were injected into the reactor using calibrated gas-tight syringes with the reactor under reduced pressure; solid compounds (nitrocresols) were introduced into the chamber in an air flow through a special heated inlet system. The initial reactant concentrations (in molecules cm-3 units) were as follows: C5H8 2× 1014, C4H10 (6-8) × 1013, nitrocresols (4.8 - 9.2) × 1013, CH3ONO 5 × 1013, C2H4 (6 – 8) × 1013 and NO 3 × 1014.

157

FT-IR Kinetic Study on the Gas-Phase Reactions of the OH Radical

6-Methyl-2-nitrophenol was prepared by addition of sodium nitrite to o-toluidine dissolved in a concentrated solution of sulphuric acid (Winzor, 1935). The rate constants were determined using the relative kinetic technique. Ethene was used as the reference compound. The OH radicals generated in the photolysis systems will react with both the nitrocresols and the reference hydrocarbon: reactant + OH

o products

(k1)

(1)

reference + OH

o products

(k2)

(2)

Under the experimental conditions, the aromatic compounds are also lost by absorption on the surface of the reactor and by photolysis. Control experiments have been performed in order to measure these loss rates: reactant (+wall loss, hQ)

o products

(k3)

(3)

No wall deposition or photolysis was observed for the reference hydrocarbons. For the systems under investigation the following kinetic expression is valid:

ln

[reactant]t o [reactant]t

- k 3 t - t o

[reference]t o k1 ln k2 [reference]t

(4)

where [reactant]t0, [reactant]t, [reference]t0 and [reference]t are the initial concentration of the aromatic compound and the concentration at time t, respectively the initial concentration of reference hydrocarbon and the concentration at time t. Hence plots of ln([reactant]t0/[reactant]t) – k3(t-t0) against ln([reference]t0/[reference]t) should yield a straight line of slope k1/k2 with zero intercept. Results The experimental kinetic data plotted in Figure 1 according to eq (4) are preliminary and are presently based on only single experiment and one reference hydrocarbon for each compound. The plots show reasonable linearity considering that the reactions are fairly slow and also the difficulties which arise in the infrared spectral subtraction procedures for these aromatic hydrocarbons. Table 1 shows the preliminary rate coefficients extracted from the plots in Figure 1 for the reaction of OH with the four nitrophenols investigated. A rate constant of 8.52 × 10-12 cm3 s-1 was used for the reaction of OH with the reference ethene (Atkinson and Arey, 2003). Photolysis of methylated nitrophenols is a potential source of OH radicals and thus also a potential source of error in the determination of the correction for photolysis in the kinetic analysis. To obtain the true values of the photolysis rates for the compounds it was necessary to perform separate experiments in which the OH radicals were scavenged. Using isoprene as scavenger it was possible to calculate the OH radical concentration and thus calculate the overestimation in the measured photolysis rate due to reaction with OH. The calculated concentrations of OH radicals produced in the photolysis of the various nitrophenols are presented in Table 2.

I. Bejan et al.

158 Table 1.

Preliminary rate constants for the reactions of OH with four i-methyl-2nitrophenols (i = 3, 4, 5, 6). The literature values (a) are computer estimates.

compound

reference

k1/k2

k1 (10 *cm3s-1)

k (literature)a (10-12*cm3s-1)

3-methyl-2nitrophenol

ethene

0.433±0.019

3.7±0.162

11.2

4-methyl-2nitrophenol

ethene

0.412±0.019

3.5±0.166

5.38

5-methyl-2nitrophenol

ethene

0.671±0.037

5.7±0.313

11.2

6-methyl-2nitrophenol

ethene

0.317±0.019

2.7±0.167

-

Table 2.

-12

Photolysis rates and the steady-state OH radical concentration estimated during the photolysis of all four i-methyl-2-nitrophenols (i = 3, 4, 5, 6) investigated. (a) values for the photolysis rate obtained in the 1080 L quartz glass reactor.

Compound

3-methyl-2nitrophenol

4-methyl-2nitrophenol

5-methyl-2nitrophenol

6-methyl-2nitrophenol

photolysis rate(a) (s-1)

1.67±0.11 x10-4

8.86±1.07x10-5

1.07±0.14x10-4

7.81±1.11x10-5

OH concentration (cm-3)

~3x105

~2x105

~3x105

~2x105

FT-IR Kinetic Study on the Gas-Phase Reactions of the OH Radical

Figure 1.

159

Kinetic data plots for the reactions of methylated-2-nitrophenols with OH radicals using ethene as reference hydrocarbon: (Ŷ) 3-methyl-2-nitrophenol (3M2N), (Ÿ) 4-methyl-2-nitrophenol (4M2N), (i) 5-methyl-2-nitrophenol (5M2N), (x) 6-methyl-2-nitrophenol. (The offsets have been introduced for clarity).

The magnitude of the photolysis rates for methylated nitrophenols under atmospheric conditions can be roughly estimated from the experiments using the superactinic lamps. This is accomplished by adjusting the values obtained in the reactor and presented in Table 2, with a factor comprised of the ratio of NO2 photolysis measured in the atmosphere and that measured in the reactor. For example, a factor of JNO2 (atmosphere)/ JNO2 (in the reactor) = (8.5 ± 0.5) x 10-3 s-1/ (2.0 ± 0.2) x 10-3 s-1 = 4.25 is obtained when using an atmospheric noontime photolysis frequency of NO2 typical for clear sky conditions at a latitude of 40o N for July 1 (Klotz et al., 1997). The loss processes (wall deposition and photolysis) make a large contribution of around 50-55% to the measured total loss of most of the compounds investigated. In the case of 5-methyl-2-nitrophenol the contribution was between 29 - 33%. This lower contribution from photolysis and wall loss to the overall loss for this compound can be explained by the higher reactivity of 5-methyl-2-nitrophenol towards OH radical compared to the other nitrocresols. The losses due solely to photolysis were approximately 23-27% in all cases.

I. Bejan et al.

160 Remarks and conclusions

This work represents the first reported experimental kinetic study of the reaction of OH radicals with methylated nitrophenols, therefore, a comparison with other experimentally obtained values is not possible. Because of the difficulties in handling the compounds the measured rate constants should be used with caution. The measured values presented here need verification by further experiments and also a reference hydrocarbon other than ethene. Validation by an independent absolute kinetic method is also desirable. Up to now there has only been a theoretical estimation of the rate constants for the atmospheric gas-phase reaction of these compounds with OH radicals (Meylan and Howard, 1993). According to the estimate both 3M2N and 5M2N should have the same rate constant of 1.12 × 10-11 cm3 s-1 and 4M2N a value of 5.4 × 10-12 cm3 s-1. 6M2N was not estimated. These results of the present work are in disagreement with the estimated values; all of the measured values are lower than the estimates with the differences being factors of 2 and 3 in the cases of 5M2N and 3M2N, respectively. The measured rate coefficients have values which are larger than that of OH + 2nitrophenol (Atkinson et al. 1992) and lower than those of OH + cresol isomers (Calvert et al. 2002). This is in line with reactivity expectations. The rate coefficient for all four nitrocresols is around 10-12 cm3 s-1, however, the preliminary results appear to indicate an influence of the position of the methyl group on the aromatic ring on the rate coefficient. Based on established chemical activating/deactivating effects for ring substituents (discussed below), the rate for OH + 5M2N would be predicted to have a higher rate than the others and OH + 6M2N should have the lowest rate constant, 3M2N and 4M2N would be expected to have fairly similar rate constants. This is roughly what is observed experimentally. The reaction of phenolic compounds with OH radicals occurs predominantly by electrophilic addition of OH to the aromatic ring (Atkinson, 2000). NO2 is a meta director because electrophilic attack at the ortho and para positions generates an especially poor resonance structure and the CH3 group is and ortho-para director. OH has a strong electron donating resonance effect that activates the ring. For all four investigated compounds NO2 and OH are fixed on adjacent positions on the aromatic ring. Thus the position of CH3 substituent has a profound affect on the reactivity of aromatic compounds with OH radicals, as can be seen in Figure 3. OH

OH NO2

*° *°

CH3

OH NO2

* °

°

NO2

*° H3C

OH H3C

NO2 °

°

*°

*

5M2N

6M2N

CH3 3M2N

Figure 3.

4M2N

The nitrocresol isomers considered in this study and the sites that are activated by the hydroxyl group (*) and methyl group (°) toward electrophilic addition.

The OH and CH3 entities both donate electron density to the aromatic ring and activate the ortho and para positions toward electrophilic addition. The directing + I effect of the methyl group is considerably less than the + E effect of OH, but nonetheless the CH3 group

FT-IR Kinetic Study on the Gas-Phase Reactions of the OH Radical

161

still plays an important role. OH activation and NO2 deactivation effects are in the 4th and 6th positions for all compounds, so additional activation from CH3 becomes important. The faster rate coefficient for the reaction of 5M2N compared with those of the other iM2Ns (i=3,4,6) can be ascribed to enhanced activation of the 4th and 6th ring positions by both the methyl and OH groups (ortho- by CH3, and both ortho- and para- by OH). A similar effect is present in 3M2N, but the + I effect of CH3 is somewhat suppressed by the – E effect of the NO2 group. Within the experimental error limits, the rate coefficients for 4M2N and 6M2N are almost the same. This is expected because the number of possible addition sites for the OH radical is the same for both compounds. Further studies are in progress to validate the rate coefficients presented here and also to determine both the photolysis and OH reaction products. References Atkinson, R. and J. Arey, Atmospheric Degradation of Volatile Organic Compounds, Chem. Rev. 103 (2003) 4605-4638. Atkinson, R., Atmospheric chemistry of VOCs and NOx, Atmos. Environ. 34 (2000) 2063-2101. Atkinson, R., S. Aschmann, and J. Arey, Reaction of OH and NO3 radicals with phenol, cresol, and 2nitrophenol at 296 ± 2 K, Environ. Sci. Technol., 26 (1992), 1397-1403 Barnes, I., K. H. Becker and N. Mihalopoulos, An FT-IR Product Study of the Photo-oxidation of Dimethyl Disulphide, J. Atmos. Chem., 18 (1994) 267-289 Bejan, I., I. Barnes, R. Olariu and R. Mocanu, Secondary organic aerosol formation from the photolysis of nitrophenols and nitrocresols, Poster presented at European Geosciences Union, 1st General Assembly, Nice, France, 25-30 April 2004 Belloli, R., B. Barletta, E. Bolzacchini, S. Meinardi, M. Orlandi and B. Rindone, Determination of toxic nitrophenols in the atmosphere by high-performance liquid chromatography, J. Chromatogr. A, 846 (1999), 277-281. Calvert, J., R. Atkinson, K.H. Becker, R. Kamens, J. Seinfeld, T. Wallington and G. Yarwood, The Mechanisms of the Atmospheric Oxidation of Aromatic Hydrocarbons, Oxford University Press, 2002. Derwent, R. G., M. E. Jenkin, S. M. Saunders and M. J. Pilling, Photochemical ozone creation potentials for organic compounds in Northwest Europe calculated with a master chemical mechanism, Atmos. Environ. 32, 14/15, (1998) 2429-2441. Derwent, R. G., M. E. Jenkin and S. M. Saunders, Photochemical ozone creation potentials for a large number of reactive hydrocarbons under european conditions, Atmos. Environ. 30, 2, (1996) 181-199. Forstner, H.J.L., R.C. Flagan and J.H. Seinfeld, Secondary organic aerosol from the photo-oxidation of aromatic hydrocarbon: Molecular composition, Environ. Sci. Techn 31, (1997) 1346-1358. Geissler, A. and H. F. Scholer, Gas chromatographic determination of phenol, methylphenols, chlorophenols, nitrophenols and nitroquinones in water. Water Research 28 (1994) 2047–2053. Grosjean, D., Atmospheric fate of toxic aromatic compound, Sci. Total Environ., 100 (1991), 367-414. Harrison, M. A. J., S. Barra, D. Borghesi, D. Vione, C. Arsene and R. I. Olariu, Nitrated phenols in the atmosphere: a review, Atmos. Environ., 39 (2005) 231-248 Herterich, R., Gas chromatographic determination of nitrophenols in atmospheric liquid water and airborne particulates. J. Chromatography, 549 (1991) 313-324. Hurley, M.D., O. Sokolov, T.J. Wallington, H. Takekawa, M. Karasawa, B. Klotz, I. Barnes and K.H. Becker, Organic Aerosol Formation During the Atmospheric Degradation of Toluene, Environ. Sci. Techn., 35 (2001) 1358-1366. Kawai, A., S. Goto, Y. Matsumoto and H. Matsushita, Mutagenicity of aliphatic and aromatic nitro compounds industrial materials and related compounds, Jap. J. Ind. Health, 29 (1987), 34-54.

162

I. Bejan et al.

Kawamura K. and I. R. Kaplan, Biogenic and anthropogenic organic compounds in rain and snow samples collected in southern California, Atmos. Environ. 20 (1986) 115–124. Klotz, B., I. Barnes, K.H. Becker and B. T. Golding, Atmospheric chemistry of benzene oxide/oxepin, J.Chem. Soc., Faraday Trans., 93 (1997) 1507-1516 Lüttke, J. and K. Levsen, Phase partitioning of phenol and nitrophenols in clouds Atmos. Environ., 31 (1997), 2649-2655. Meylan, W. M. and P. H. Howard, Computer estimation of the atmospheric gas-phase reaction rate of organic compounds with hydroxyl radicals and ozone, Chemosphere, 26, 12 (1993) 2293-2299. Nojima, K., K. Fukaya, S. Fukui and S. Kanno, The formation of nitrophenol and nitrobenzene by the photochemical reaction of benzene in the presence of nitrogen monoxide, Chemosphere, 2 (1975), 77-82. Odum, J.R., T. Hoffmann, F. Bowman, D. Collins, R.C. Flagan and J.H. Seinfeld, Gas/particle partitioning and secondary organic aerosol yields, Environ. Sci. Techn. 30 (1996) 2580-2585. Olariu, R. I., B. Klotz, I. Barnes, K. H. Becker and R. Mocanu, FT-IR study of the ring- retaining products from the reaction of OH radicals with phenol, o-, m- and p-cresol, Atmos. Environ., 36 (2002) 3685-3697 Tremp, J., P. Mattrel, S. Fingler, W. Giger, Phenol and nitrophenols as tropospheric pollutants: emissions from automobile exhausts and phase transfer in the atmosphere. Water Air Soil Pollut., 68 (1993), 113-123. Voznakova, Z., J. Podehradska, and M. Kohlickova, Determination of nitrophenol in soil, Chemosphere, 33, 2 (1996), 285-291 Winzor, F. L., The colouring matters of Drosera Whittakeri. Part III. The synthesis of hydroxydroserone. J. Chem. Soc. Lond. (1935) 336-338

Atmospheric Fate of Unsaturated Ethers A. Mellouki LCSR/CNRS, 1C Avenue de la recherche scientifique, 45071 Orléans cedex 02, France Key Words: Unsaturated Ethers, Tropospheric degradation, Kinetics, OH radicals, NO3 radicals, Ozone

Introduction Oxygenated volatile organic compounds form a major component of the trace gases found in the troposphere (Singh et al., 2001). They are emitted directly into the atmosphere from biogenic sources, solvent and fuel additives use and are also formed in the tropospheric oxidation of all hydrocarbons. A number of oxygenated compounds are used as alternatives to aromatic and halocarbon solvents both in terms of toxicity problems and in the case of aromatic compounds as a means of reducing the levels of oxidant formation in the troposphere. It is apparent that oxygenated compounds will thus play an increasing role in determining the oxidising capacity of the troposphere both on a regional and global scale. The degradation of oxygenated volatile organic compounds in the troposphere leads to the production of a range of secondary pollutants such as ozone, nitrates and secondary organic aerosols. Ozone formation is of particular concern since it is known to have adverse effects on human health, vegetation and a variety of materials, and it is also a greenhouse gas. Indeed, strategies for the control of emissions of volatile organic compounds have been based on their individual contribution to the photochemical formation of ozone (Derwent et al., 1996). Information on the fate of oxygenated compounds in the troposphere is not only important in terms of oxidant formation, but oxidation of simple oxygenated compounds often leads to the generation of poly-functionalized species. These compounds can be of low reactivity in the gas phase and thus their relatively high solubility and low vapour pressures may lead to uptake in cloud and fog droplets or to formation of secondary organic aerosols. Hence, anthropogenic emissions of oxygenates may impact on wet deposition and also on climate by increasing the levels of tropospheric aerosol. Because of their high reactivity, vinyl ethers can be used in a variety of syntheses: e.g. polymerizations, addition, and electrocyclic reactions. They are used in different industries, particularly as solvents, motor oil additives, for the manufacturing of coatings or as intermediates for the synthesis of flavours, fragrances and pharmaceuticals. This broad utility has made many applications possible, and is expected to create even more applications in the future (http://www.basf.de). In the atmosphere, the degradation of unsaturated oxygenates is controlled mainly by chemical reaction with OH and NO3 radicals and O3 while photolysis plays a negligible role. A number of studies have been conducted recently, within an EU project to investigate the atmospheric fate of the unsaturated ethers (Georges et al., 2004). In this contribution, some of the experimental data obtained on the gas phase processes are presented. This includes kinetic and mechanistic data for the reactions of OH, O3 and NO3 with a series of vinyl ethers (methyl vinyl ether (MVE) CH3OCH=CH2, ethyl vinyl ether (EVE) C2H5OCH=CH2, propyl vinyl ether (PVE) n-C3H7OCH=CH2, butyl vinyl ether (BVE) n-C4H9OCH=CH2, iso-butyl vinyl ether (IBVE) i-C4H9OCH=CH2 and tert-butyl vinyl ether (TBVE) t-C4H9OCH=CH2). 163 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 163–169. © 2006 Springer. Printed in the Netherlands.

A. Mellouki

164 OH-initiated oxidation of vinyl ethers

The rate constants of the reactions of OH with the vinyl ethers have been measured in the temperature and pressure ranges 230-373 K and 30-300 Torr using the pulsed laser photolysis-laser induced fluorescence (PLP – LIF) method and also the relative kinetic method at 760 Torr and 298 K. The measured rate coefficients in the temperature range 230372 K are shown in the conventional Arrhenius form (k = Ae-Ea/RT) in Figure 1. The data obtained at room temperature as well as the Arrhenius parameters are summarised in Table 1. The Arrhenius parameters derived from the temperature dependence study show a negative temperature dependence for all measured rate constants which supports a mechanism proceeding mainly by addition of OH to the double bond of the vinyl ether.

Figure 1.

Rate constants for the reactions of OH with the studied vinyl ethers in the temperature range 230-373 K plotted according to the Arrhenius expression. The solid lines represent the Arrhenius parameter least-squares fit to the individual data points. The error bars of the individual points are 2V and do not include systematic errors (Thiault et al., 2002 and 2005).

The rate constant values of the OH reaction with vinyl ethers are higher than those with simple ethers (ROR’, with R, R’ = CH3, C2H5, C3H7, C4H9), in the ranges (4 – 11) × 10-11 and (0.3 – 3) × 10-11 cm3 molecule-1 s-1, respectively. This is in line with an OH-addition mechanism to the double bond as in the reaction of OH with alkenes. However, the values of the OH reactions with vinyl ethers, as shown in Table 2, are up to three times higher than those of OH with the corresponding alkenes. This would suggest that –OR groups in the vinyl ethers activate the double bond by electron donation more than alkyl groups –R do. In addition, while the reactivity of OH with alkenes does not seem to be influenced by the nature of the R group (k(OH+alkene) | 3×10-11 cm3 molecule-1 s-1) with R = CH3 to C4H9, that of OH

Atmospheric Fate of Unsaturated Ethers

165

with vinyl ethers increases from 4×10-11 for CH3OCH=CH2 (Perry et al., 1977) to 1.1×10-10 cm3 molecule-1 s-1 for C3H7OCH=CH2 and C4H9OCH=CH2. Table 1.

Summary of the measured rate constants at 298 K obtained by absolute and relative methods and the Arrhenius parameters in the temperature range 230372 K. a)

Vinyl ether

a)

a)

A u 10

11

b)

k298 K u 1010

a)

k298 K u 1010

E/R Absolute

Relative

MVE, CH3OCH=CH2

-

-

-

0.45 r 0.07

EVE, C2H5OCH=CH2

1.55 r 0.25

-(445 r 13)

0.68 r 0.07

0.72 r 0.08

PVE, C3H7OCH=CH2

0.93 r 0.06

-(708 r 20)

1.0 r 0.1

1.1 r 0.1

BVE, n-C4H9OCH=CH2

1.48 r 0.22

-(572 r 42)

1.0 r 0.1

1.1 r 0.1

IBVE, i-C4H9OCH=CH2

1.57 r 0.11

-(567 r 20)

1.1 r 0.1

1.1 r 0.1

TBVE, t-C4H9OCH=CH2

1.70 r 0.15

-(549 r 25)

1.1 r 0.1

1.1 r 0.1

in cm3 molecule-1 s-1 ; b) in K.

Table 2.

Comparison of the rate constant values for the reaction of OH radicals with vinyl ethers and alkenes at 298 K (k in cm3 molecule-1 s-1).

Alkyl group (R)

Alkenes (RCH=CH2)

Vinyl Ethers (ROCH=CH2)

CH3

2.63 x 10-11

3.4 x 10-11

C2 H5

3.14 x 10-11

7 x 10-11

n-C3H7

3.14 x 10-11

10 x 10-11

n-C4H9

3.7 x 10-11

11 x 10-11

i-C4H9

3.8 x 10-11

11 x 10-11

t-C4H9

2.8 x 10-11

11.2 x 10-11

As shown in Table 3, the structure reactivity relationships (SAR) developed by Atkinson et al. does not apply to the vinyl ethers since the calculated rate constant values are lower than the measured ones for the vinyl ethers studied in this work. The main observed products of the OH-initiated oxidation of vinyl ethers in the presence of NOx, were formaldehyde and formates (with molar yields of | 80 % for each).

A. Mellouki

166

The observed products can be explained by a mechanism proceeding mainly by OH-radical addition to the carbon-carbon double bond: ROCH=CH2 + OH (+O2 + NO + M) o ROCH(OH)CH2O° + NO2

(1a)

ROCH=CH2 + OH (+O2 + NO + M) o ROCH(O°)CH2OH + NO2

(1b)

The hydroxyalkoxy radicals ROCH(OH)CH2O° and ROCH(O°)CH2OH decompose, react with O2, or isomerise. The reactions of these two radicals with O2 may lead to hydroxycarbonyl compounds (ROCH(OH)CHO and ROC(O)CH2OH) while their decomposition will lead to the corresponding formate and formaldehyde: ROCH(OH)CH2O° (+M) o ROC°HOH + HCHO ROC°HOH + O2 o ROC(O)H + HO2 and ROCH(O)CH2OH (+M) o ROC(O)H + CH2OH CH2OH + O2 o HCHO + HO2 Table 3.

Comparison between the measured rate constant values for the reaction of OH with vinyl ethers and the calculated values using the SAR of Atkinson et al. (Atkinson, 1987; Kwok and Atkinson, 1995).

Vinyl ether

kcalculated

kmeasured

(cm3 molecule-1 s-1)

(cm3 molecule-1 s-1)

CH3OCH=CH2

3.5 x 10-11

3.4 x 10-11 (Perry et al., 1977)

C2H5OCH=CH2

4.2 x 10-11

7 x 10-11

n-C3H7OCH=CH2

4.5 x 10-11

10 x 10-11

n-C4H9OCH=CH2

4.7 x 10-11

11 x 10-11

i-C4H9OCH=CH2

4.7 x 10-11

11 x 10-11

t-C4H9OCH=CH2

3.5 x 10-11

11 x 10-11

The detection of high yields of formates and formaldehyde indicates that the decomposition of the two hydroxyalkoxy radicals dominates over the other processes (reaction with O2 and isomerisation). The other possible mechanism is H-atom abstraction from the R groups which produces vinyl esters such as R’C(O)OCH=CH2 and HC(O)OCH=CH2 when the abstraction is from the –CH2- group attached to –O- entity: R’CH2OCH=CH2 + OH (+O2 + NO + M) o R’CH(O°)OCH=CH2 + NO2 R’CH(O°)OCH=CH2 (+O2) o R’C(O)OCH=CH2 R’CH(O°)OCH=CH2 (+M +O2) o HCHO + H(O)COCH=CH2

Atmospheric Fate of Unsaturated Ethers

167

An H-abstraction from the R’ group will lead to other species, however, the only products observed were formaldehyde and formates which indicates that the OH-radical oxidation of vinyl ethers proceeds essentially by OH addition to the carbon-carbon double bond. O3 Reactions with Vinyl Ethers The measured values for the rate constants of the reaction of O3 with vinyl ethers are presented in Table 4. The are in the range (2-5)×10-16 cm3 molecule-1 s-1, i.e. they are higher than those for simple alkenes again indicating that the presence of the O atom also in this case enhances the reactivity with the C=C double bond. It would appear that the nature of the R group in the vinyl ethers (ROCH=CH2) also has an effect on the reactivity. As in the OH-initiated oxidation of vinyl ether, the reactions of O3 with the same vinyl ethers lead mainly to formation of formates (molar yields of (70-90)%) and formaldehyde (molar yields of (20-10)%). Figure 2 shows an example of the concentration-time profiles of the reactants and products for the reaction of O3 with ethyl vinyl ether (C2H5OCH=CH2). For methyl vinyl ether, hydroperoxymethyl formate and methanol were also observed as products (Klotz et al., 2004). Table 4.

O3 + vinyl ethers, summary of the measured rate constants at 298 ± 2 K. Vinyl Ethers

k (cm3 molecule-1 s-1)

EVE, C2H5OCH=CH2

(2.0 r 0.2) u 10-16

PVE, C3H7OCH=CH2

(2.4 r 0.4) u 10-16

BVE, n-C4H9OCH=CH2

(2.9 r 0.2) u 10-16

IBVE, (CH3)2CHCH2OCH=CH2

(3.1 r 0.2) u 10-16

TBVE, (CH3)3COCH=CH2

(5.0 r 0.5) u 10-16

As for the reactions of O3 with olefins, the reaction of O3 with ROCH=CH2 proceeds by initial addition of O3 to the carbon-carbon double bond, followed by decomposition of the ozonide to form formaldehyde and a formate as stable compounds in addition to two Criegee intermediates: O3 + ROCH=CH2 ĺ ROC°HOO° + HCHO O3 + ROCH=CH2 ĺ ROC(O)H + HC°HOO° The energy rich Criegee intermediates ROC°HOO° and HC°HOO° may thermally stabilize, rearrange, react or decompose to form other stable compounds. Laboratory studies have shown that the O3 initiated oxidation of vinyl ethers may lead to the formation of secondary organic aerosols. The reported mass based formation yields are in the range (0.5-10) % depending on the starting organic compound (Klotz et al., 2004; Georges et al., 2004).

A. Mellouki

168

EVE Ethyl Formate HCHO Ozone

EVE concentration (ppb)

800

200

700

100 600

0

500

0

Figure 2.

Products and ozone concentrations (ppb)

300

900

5

10 15 Reaction time (min)

20

25

Concentrations of products and reactants versus reaction time in the O3 initiated oxidation of C2H5OCH=CH2 in air (Thiault et al., 2002).

NO3 Reaction with vinyl ethers Kinetic and mechanistic investigations on the NO3 radical oxidation of three vinyl ethers, methyl, propyl and butyl vinyl ethers have been performed in smog chambers (Klotz et al., 2004, Georges et al., 2004). Analyses were performed using an in situ FTIR device coupled to a multi-reflection White-cell. Rate constants for the oxidation of methyl and propyl vinyl ethers with NO3 were determined using the relative rate technique and/or by a “pseudoabsolute” rate technique. Rate constants of (4 r 2) ×10-13 and (1.3 r 0.2) ×10-12 cm3 molecule1 -1 s were obtained for the NO3 oxidation of methyl and propyl vinyl ethers, respectively. Mechanistic experiments also have been performed for methyl, propyl and butyl vinyl ethers. Main oxidation products are formaldehyde, a formate (methyl, propyl or butyl), nitrates and peroxynitrates (Figure 3). Formaldehyde and the formates have been quantified and their formation yields estimated to be around 50 %. The reaction of NO3 with unsaturated ethers proceeds by addition to the >C=C< double bond. Both steric and inductive effects are known to influence the reactivity of NO3 with unsaturated alcohols, e.g. 2-methyl-3-butene-2-ol is slightly less reactive than 1-butene. Unsaturated ethers are more reactive than unsaturated alcohols towards NO3. Comparison of the measured rate constants shows that, similar to the reaction with O3, the presence of the carbonyl group decreases the reactivity of NO3 toward the >C=C< double bond. The reaction of NO3 with unsaturated oxygenates proceeds by addition to either one of the carbon atoms of >C=C< double bond, preferentially to the most substituted radical. In the presence of NOx, the reaction may ultimately lead to organic nitrates among the first generation products.

Atmospheric Fate of Unsaturated Ethers

169

3

PVE + NO3 PVE

2.5

Propyl Formate

HCHO

Concentration (ppm)

N2O5

HNO3

2

NO2 Sum Nitrates (arbitrary units)

1.5

1

0.5

0 0

500

1000

1500

2000

2500

3000

3500

Time (s)

Figure 3.

Concentration profiles for an experiment of NO3-induced oxidation of propyl vinyl ether (Georges et al., 2004).

Atmospheric Implications In the atmosphere, unsaturated ethers are oxidised in the gas phase by reaction with OH, NO3 and O3. The tropospheric lifetimes of unsaturated ethers are typically of few hours and oxidation results mainly in the production of aldehydes and formates which are less reactive than their parent compounds. It has been found that the reactions of ozone with vinyl ethers lead to the formation of secondary organic aerosols under laboratory conditions. This aerosol formation warrants further investigation in order to assess its importance under real atmospheric conditions. References Atkinson R.; A structure activity relationship for the estimation of rate constants for the gas-phase reactions of OH radicals with organic compounds, Int. J. Chem. Kinet. 19 (1987) 799-828. Derwent, R. G.; Jenkin, M. E.; Saunders, S. M. Atmos. Environ. 30 (1996) 181. Georges, C. et al., Multiphase chemistry of Oxygenated Species in the Troposphere – MOST, EVK2-CT-200100114 (2002-2004). Klotz B., I. Barnes and T. Imamura; Product study of the gas-phase reactions of O3, OH and NO3 radicals with methyl vinyl ether, Phys. Chem. Chem. Phys. 6 (2004) 1725-1734. Kwok E. S. C. and R. Atkinson; estimation of hydroxyl radical reaction rate constants for gas phase organic compounds using a structure-reactivity relationship: an update, Atmos. Environ. 29 (1995) 1685-1695. Perry R.A., R. Atkinson, J. N. Pitts; Rate Constants for the Reaction of OH Radicals with Dimethyl Ether and Vinyl Methyl Ether over the Temperature Range 299-427oK, J. Chem. Phys., 67 (1977) 611. Singh, H.; Chen, Y.; Standt, A.; Jacob, D.; Blake, D.; Heikes, B.; Snow, J. Nature 410 (2001) 1078-1081. Thiault G., R. Thévenet, A. Mellouki, G. Le Bras; OH and O3-initiated oxidation of ethyl vinyl ether Phys. Chem. Chem. Phys. 4 (2002) 613-619. Thiault G., A. Mellouki, G. Le Bras ; Rate constants for the OH radical reaction with n-propyl, n-butyl, iso-butyl and tert-butyl vinyl ethers to be submitted (2005).

Atmospheric Oxidation of the Chlorinated Solvents, 1,1,1-Trichloroethane, Trichloroethene and Tetrachloroethene Lorraine Nolan, Anne-Laure Guihur, Marcus Manning, and Howard Sidebottom Department of Chemistry, University College Dublin, Belfield, Dublin 4, Ireland Key Words: Chlorinated solvents, Chloroalkoxy radicals, Photolysis, Troposphere, Hydroxyl radical

Introduction Chlorinated solvents are widely used in metal degreasing, dry cleaning, paint stripping, extraction of pharmaceuticals and foodstuffs, and electronic circuit board production (Sidebottom and Franklin, 1996). Emission data for trichloroethene (CHCl=CCl2) and tetrachloroethene (CCl2=CCl2) show that emissions of these compounds are declining steadily, largely as a result of constant improvements in the efficiency with which they are being used and recycled. While 1,1,1-trichloroethane (CH3CCl3) was formerly widely used as a solvent, it is now strictly regulated as an ozone-depleting compound under the Montreal Agreement. Despite the decline in emissions of these chlorinated solvents their atmospheric fate and impact is still of concern. Chloral (CCl3C(O)H) is the major product from the OH radical initiated oxidation of CH3CCl3 (Nelson et al., 1990, Platz et al., 1995). The available data show that the tropospheric lifetime of CCl3C(O)H is of the order of a few hours and is predominantly controlled by photolysis (Talukdar et al., 2001, Wenger et al., 2004). It has been shown that a chlorine atom is produced in the primary photolysis process at 308 nm with a quantum yield close to unity and it was assumed that the HC(O)CCl2 radical, which is also formed in the photolysis, generates phosgene (Cl2C(O)) in the presence of O2 (Talukdar et al., 2001). Recently, the photolysis of chloral under atmospheric conditions has been investigated at the large outdoor European photoreactor in Valencia, Spain (Wenger et al., 2004). The photolytic lifetime of chloral was shown to be 4.5-6 hours under conditions appropriate to the solar flux during summer months and the quantum efficiency of photolysis under atmospheric conditions was determined to be 1.00 ± 0.05. The reported atmospheric photolysis product yield data are consistent with a mechanism in which the primary process generates a Cl atom and a HC(O)CCl2 radical. The HC(O)CCl2 radical forms the corresponding alkoxy radical 3HC(O)CCl2O which decomposes by both C-C (~85%) and C-Cl (~15%) bond fission. This work is concerned with reactions of the alkoxy radicals HC(O)CCl2O and ClC(O)CCl2O under atmospheric conditions in order to determine the relative importance of the carbon-carbon and carbon-chlorine bond fission pathways for their degradation. The OH radical initiated oxidations of the chloroethenes CHCl=CCl2 and CCl2=CCl2 in the presence of a Cl atom scavenger have been reported to give Cl2C(O) as a major product with yields of around 50% for both alkenes (Tuazon et al., 1988). The relatively low yields of Cl2C(O) and the lack of HClC(O) as a product in the oxidation of CHCl=CCl2 suggests that the energy-rich alkene adduct, formed in the reactions, may eliminate HCl rather than add molecular oxygen. The resulting radicals will lead to the alkoxy radicals HC(O)CCl2O and ClC(O)CCl2O. Hence the breakdown pathways of these radicals may be of importance in understanding the atmospheric oxidation of tri-and tetrachloroethene. 171 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 171–179. © 2006 Springer. Printed in the Netherlands.

L. Nolan et al.

172 Experimental

Product distribution studies in the laboratory were carried out at 298 ± 2K and atmospheric pressure in a FEP Teflon chamber with a volume of approximately 50 L. The temperature of the reaction chamber was thermostatically controlled and electric fans positioned below the chamber helped maintain a constant temperature. The reaction chamber was surrounded by banks of fluorescent lamps, which provided sources of radiation at 254 nm (Philips germicidal lamps TUV, 15 W) and in the 300 to 450 nm wavelength region with a maximum intensity around 365.5 nm (Philips black lamps TL 20/08). The reaction chamber was approximately half-filled by a diluent gas, typically zero-grade air, via Teflon tubing using calibrated flow meters. Measured pressures (MKS Baratron capacitance manometers) of the reactants were then expanded into the evacuated vacuum line and isolated in a calibrated mixing bulb, from which they were swept into the reaction chamber by a stream of zero-grade air. Hydrogen peroxide solution was injected directly into the chamber. When all the reactants had been added to the chamber it was subsequently filled to capacity at a pressure of 1 atmosphere and kept in the dark for 1 hour to allow complete mixing of the reactants. A homogeneous reaction mixture was confirmed by consistent, reproducible gas chromatographic analysis. Hydroxyl radicals were generated by the photolysis of hydrogen peroxide. H2O2 + hȞ (Ȝ = 254 nm)

ĺ

OH + OH

(1)

Molecular chlorine was photolysed to produce chlorine atoms. Cl2 + hȞ (Ȝ > 300 nm)

ĺ

Cl + Cl

(2)

Quantitative analyses of the reaction mixtures were carried out by gas chromatography (Shimadzu GC 14A coupled with a flame ionisation detection), and Fourier transform infrared spectroscopy (Mattson Research Series 1 fitted with a deuterated triglycine sulfate (DTGS) detector). The spectrometer was equipped with an evacuable 5 L Teflon–coated Wilks cell, containing a multipass White mirror arrangement (pathlength 7-10 m) to facilitate the collection of infrared spectra. Spectra were recorded over the range 400-4000 cm-1 and with a resolution of 2 cm-1. Each resulting spectrum was derived from the co-addition of 20 scans under these conditions. Authentic samples of reactants and products were used to calibrate reference spectra. The reaction chamber employed for product studies at the EUPHORE facility consisted of a half-spherical FEP Teflon chamber with a volume of around 195,000 L mounted on aluminium floor panels covered with FEP Teflon foil. The FEP Teflon had a transmission of more than 80% over the wavelength range 280-640 nm. A hydraulically operated steel housing surrounded the chamber and allowed the exposure time to sunlight to be controlled. Heating of the reaction mixtures above ambient temperature was compensated with cooling by a refrigeration system housed in the chamber floor. Typically, the chamber temperature was in the range 300 ± 5 K. The chamber was filled with air from the air purification system consisting of absorption dryers connected to charcoal scrubbers. Two mixing fans housed in the chamber were used to ensure homogenous reaction mixtures. Measured volumes of gaseous and liquid samples were introduced into the covered reaction chamber through Teflon tubing by a stream of purified air. Losses of reactants and products occurred by leakage from the chamber, which was maintained at several Torr above ambient pressure. The leak rate was determined by adding an unreactive tracer gas, SF6 to the chamber. The loss of reactants and formation of products was monitored with a Nicolet

Atmospheric Oxidation of the Chlorinated Solvents

173

Magma 550 FTIR spectrometer equipped with a mercury cadmium telluride (MCT) detector positioned on a platform beneath the chamber. Infrared spectra were obtained by in situ longpath absorption, and were derived from the co-addition of 900 scans recorded by using a path length of 553.5 m over the range 400-4000 cm-1 with a resolution of 1 cm-1. The reactants were quantified by using calibrated reference spectra. Results and discussion The photolysis of chloral (50-100 ppm) in air at Ȝ > 300 nm was carried out in the laboratory in the presence of excess isopropanol (250-500 ppm) as a chlorine atom trap. The major products from the reaction were Cl2C(O) and CO with yields of 0.87 ± 0.03 and 1.01 ± 0.03 respectively. The loss of the added isopropanol (1.28 ± 0.06) and consequent formation of acetone (1.30 ± 0.05) provide a measure of the yield of chlorine atoms in the system. Cl + (CH3)2CHOH

ĺ

(CH3)2COH + HCl

(3)

(CH3)2COH + O2

ĺ

HO2 + (CH3)2C(O)

(4)

A similar series of photolysis experiments was carried out at the European outdoor reaction chamber with initial concentrations of chloral and isopropanol in the range 0.4-1.2 ppm. The measured yields of Cl2C(O) (0.85 ± 0.01), CO (1.05 ± 0.05), (CH3)2CHOH (-1.39 ± 0.03) and (CH3)2C(O) (1.28 ± 0.03) were in agreement with the experimental uncertainties determined in the laboratory system. Assuming the photolysis of chloral at Ȝ > 300 nm yields a chlorine atom and a HC(O)CCl2 radical (Talukdar et al., 2001), the yield of Cl2C(O) from the photolysis of chloral indicates that the HC(O)CCl2 radical is not quantitatively converted to Cl2C(O). Thus, decomposition of the HC(O)CCl2O radical, generated from HC(O)CCl2, occurs by either carbon-carbon or carbon-chlorine bond fission (Figure 1). According to the mechanism given in Figure 1, C-C bond cleavage of the HC(O)CCl2O radical will generate Cl2C(O) with a yield of ~0.85 and leads to the formation of CO (~0.85) following the reaction of HC(O) with O2. The remaining ~15% of the HC(O)CCl2O radicals undergo C-Cl bond fission to produce chloroglyoxal HC(O)C(O)Cl and a chlorine atom. This proposed mechanism predicts molar yields of Cl2C(O) (~0.85), CO (~0.85), HC(O)C(O)Cl (~0.15) and Cl (~1.15) as compared to the observed yields of Cl2C(O) (~0.85), CO (~1.0) and Cl (~1.35). Chloroglyoxal was not observed among the products, however it is likely that this compound undergoes fairly rapid photolysis. It is proposed that HC(O)C(O)Cl degrades during the relatively long (~5 hours) chloral photolysis experiments to form the additional amounts of CO and Cl atoms observed. It is interesting to note that the yield of phosgene in experiments carried out in the presence of NO was virtually unchanged from the yields observed in the absence of NO. The HC(O)CCl2O radical produced from the reaction of HC(O)CCl2O2 with NO is vibrationally excited. However, the relative importance of the two unimolecular decomposition channels are almost the same as in the decomposition of HC(O)CCl2O radicals generated by the thermoneutral combination of the corresponding peroxy radicals.

L. Nolan et al.

174 CCl3C(O)H + hQ (O~ 300 nm) HC(O)CCl2 + O2 + M

HC(O)CCl2 + Cl HC(O)CCl2O2 + M

X2 / NO

NO2

HC(O)CCl2O ' (C-C)

' (C-Cl) ~15%

~85%

HC(O)C(O)Cl +Cl

HC(O) + Cl2C(O) O2

hQ

HO2 +CO

HC(O) + ClC(O) O2

HO2 + CO Figure 1.

Cl + CO

Mechanism for the photolysis of CCl3C(O)H in air.

In an attempt to further investigate reactions of alkoxy radicals of general structure RC(O)CCl2O, the chlorine atom initiated oxidation of CHCl2C(O)Cl (50-100 ppm) was investigated in air at 298 K. Product distribution studies show Cl2C(O) (0.90 ± 0.03) and ClC(O)C(O)Cl (0.11 ± 0.01) as the major products. Hydrogen atom abstraction from the methyl group in dichloroacetyl chloride leads to the formation of ClC(O)CCl2 radicals which, following addition to molecular oxygen and reaction with another peroxy radical, generate the ClC(O)CCl2O radical (Figure 2). The yields of Cl2C(O) and ClC(O)C(O)Cl indicate that the ClC(O)CCl2O radicals decay both by the C-C and C-Cl bond cleavage in a similar manner to that suggested for the structurally similar radical HC(O)CCl2O. Since the Cl2 photolysis experiments are relatively rapid (~15 minutes) the dicarbonyl ClC(O)C(O)Cl is photochemically stable over the time period of the experiment.

Atmospheric Oxidation of the Chlorinated Solvents

175

Cl2 + hQ (O > 300 nm)

Cl + Cl

Cl +CHCl2C(O)Cl

Cl(O)CCl2 +HCl

ClC(O)CCl2 + O2 + M

ClC(O)CCl2O2 + M

X2

ClC(O)CCl2O ' (C-C) ~90%

ClC(O) + Cl2C(O)

'(C-Cl) ~10%

ClC(O)C(O)Cl + Cl

Cl + CO

Figure 2.

Mechanism for the chlorine atom initiated oxidation of CHCl2C(O)Cl.

The hydroxyl radical initiated oxidations of trichloroethene (50-100 ppm) and tetrachloroethene (50-100 ppm) were investigated in air at 1 atmosphere total pressure and 298 K in the laboratory reactor. Photolysis of hydrogen peroxide (100-500 ppm) at 254 nm was used as the OH radical source. Reactions were carried out both in the absence and presence of dimethyl ether, which was used as a chlorine atom scavenger. The products found from the reactions of OH with CHCl=CCl2 and CCl2=CCl2 in the absence of CH3OCH3 show the presence of CHCl2C(O)Cl and CCl3C(O)Cl respectively, formed in the Cl atom initiated oxidation of the alkenes. Under conditions where sufficient CH3OCH3 had been added to the reaction mixtures so that the yields of acid chlorides were negligible, it was assumed that the Cl atoms generated in the OH reactions were totally scavenged. The major products detected in the OH radical initiated oxidations of CHCl=CCl2 and CCl2=CCl2 for reactions performed in the presence of sufficient CH3OCH3 to sufficiently quench the Cl atom reactions with the chloroalkenes were Cl2C(O) and CO. The yields of Cl2C(O) in both systems were approximately 70% in reasonable agreement with previously reported data (Tuazon et al., 1988). The yield of HClC(O) from the reaction of OH with CHCl=CCl2 was shown to be less than 5%. The OH radical initiated oxidations of the chloroethenes proceed by initial OH addition to the carbon-carbon double bond. All the available evidence suggests that for

L. Nolan et al.

176

unsymmetric alkenes, OH addition to the least substituted carbon atom dominates and hence reaction with CHCl=CCl2 is expected to involve the addition to the =CHCl group (Atkinson, 1994). While it is known at low pressures the hydroxyalkyl radical formed may unimolecularly decay by loss of OH, it is generally assumed that under atmospheric conditions the hydroxyalkyl formed will rapidly add O2 (Atkinson, 1994). Reaction with CHCl=CCl2 and CCl2=CCl2 can potentially lead to a complex series of reactions since the OH radical adds to a carbon atom that is bonded to a chlorine atom. The adduct radical formed by addition of OH to the alkene may eliminate a Cl atom or HCl (Figures 3 and 4).

HOCHCl-CCl2

OH + CHCl=CCl2

' (C-Cl) O / X2

HOCH=CCl2 + Cl

HC(O)CCl2 +HCl

HOCHClCCl2O ' (C-C)

CHCl2C(O)H

O / X2

HC(O)CCl2O

HOCHCl +Cl2C(O) ' (C-C)

O2

HO2 + HClC(O)

HC(O) + Cl2C(O)

' (C-Cl)

HC(O)C(O)Cl +Cl

O2

HO2 + CO

Figure 3.

Mechanism for the hydroxyl radical initiated oxidation of CHCl=CCl2.

Atmospheric Oxidation of the Chlorinated Solvents

177

HOCCl2-CCl2

OH + CCl2=CCl2

' (C-Cl) O / X2

HOCCl=CCl2 + Cl

ClC(O)CCl2 +HCl

HOCHCl2-CCl2O ' (C-C)

CHCl2C(O)Cl

O / X2

ClCC(O)CCl2O

HOCCl2 +Cl2C(O) ' (C-C)

O2

HO2 + Cl2C(O)

ClC(O) + Cl2C(O)

' (C-Cl)

ClC(O)C(O)Cl +Cl

O2

Cl + CO

Figure 4.

Mechanism for the hydroxyl radical initiated oxidation of CCl2=CCl2.

The absence of CHCl2C(O)H and CHCl2C(O)Cl from the reaction products of the OH initiated oxidation of CHCl=CCl2 and CCl2=CCl2 respectively indicates that Cl atom elimination from the OH-chloroalkene adducts is of minor importance in 1 atmosphere of air. Addition of OH radicals to CHCl=CCl2 and CCl2=CCl2 gave Cl2C(O) and CO as the major oxidation products. The relatively low yield of Cl2C(O) in the OH initiated oxidation of CCl2=CCl2 and formation of Cl2C(O) but the absence of HClC(O) in the OH initiated oxidation of CHCl=CCl2 provides evidence that elimination of HCl occurs from the initially formed energy rich adducts in these reactions. Thus, the results are consistent with the formation of HC(O)CCl2 and ClC(O)CCl2 radicals in the reactions of OH with CHCl=CCl2 and CCl2=CCl2 (Figures 3 and 4 respectively). These radicals are identical to the radicals formed by photolysis of chloral and the reaction of Cl atoms with CHCl2C(O)Cl. The absence of HClC(O) and the relatively low yield of Cl2C(O) in the CHCl=CCl2 and CCl2=CCl2 systems suggest that addition of OH to the chloroalkenes leads mainly to loss of HCl rather than addition of the adduct radicals to molecular oxygen.

L. Nolan et al.

178

Rather interestingly further information on the mechanism for the OH radical initiated oxidation of CHCl=CCl2 comes from studies on the reactions of Cl atoms with 2,2,2trichloroethanol in the presence of O2. Reaction of Cl atoms with CCl3CH2OH involves Hatom abstraction from the –CH2- group to give the CCl3CHOH radical. Reaction of this radical with O2 is expected to yield CCl3C(O)H, Figure 5. Formyl chloride and phosgene were also identified as primary oxidation products and it is suggested that these products arise from a 1,2 chlorine atom rearrangement of CCl3CHOH to give the HOCHClCCl2 radical. Reaction of this radical with molecular oxygen followed by self reaction of the peroxy radicals will generate the ȕ–hydroxy radical HOCHClCCl2O. The initial yields of CCl3C(O)H (~0.5), Cl2C(O) (~0.4) and HClC(O) (~0.4) from reactions carried out in air at 1 atmosphere suggests that under these conditions around 50% of the alkyl radicals CCl3CHOH, formed by H-atom abstraction from the parent alcohol, react with O2 to form CCl3C(O)H while the 1,2 chlorine atom shift process accounts for the balance of the products. As expected the yield of CCl3C(O)H is oxygen dependent, increasing with the oxygen pressure. It is interesting to note that HOCHClCCl2 radicals formed by 1,2 chlorine atom rearrangement are also produced by the addition of OH to CHCl=CCl2. As discussed previously, the energised adduct radical formed by addition of OH to CHCl=CCl2 appears to eliminate HCl leading to the production of Cl2C(O) as the only stable product. However, the major reaction channel for the thermalised HOCHClCCl2O radical generated from the reaction of Cl with CCl3CH2OH appears to be carbon-carbon bond fission, and both HClC(O) and Cl2C(O) are important reaction products. The proposed reaction mechanism for the Cl atom initiated oxidation of CCl3CH2OH is shown in Figure 5. Cl + CCl3CH2OH

CCl3CHOH + HCl

O2

HO2 + CCl3C(O)H

1,2-Cl atom shift

HOCHClCCl2

O2 / X2

HOCHCl + Cl2C(O)

HOCHClCCl2O

O2

HO2 + HClC(O)

Figure 5.

Mechanism for the chlorine atom initiated oxidation of CCl3CH2OH.

Atmospheric Oxidation of the Chlorinated Solvents

179

Conclusions and atmospheric implications The major reaction pathway for HC(O)CCl2O and ClC(O)Cl2O radicals, generated from the photolysis of chloral and the Cl-atom initiated of dichloroacetyl chloride respectively is decomposition via C-C bond fission to produce phosgene. These alkoxy radicals are also proposed to be intermediates in the OH radical initiated oxidation of trichloroethene and tetrachloroethene under atmospheric conditions. Thus, the major oxidation product of 1,1,1trichloroethane, trichloroethene and tetrachloroethene is Cl2C(O). Modelling study showed that tropospheric removal of Cl2C(O) is predominantly via wet decomposition with a lifetime of around 70 days (Kindler et al., 1995). Hence Cl2C(O) may provide a small flux of Cl into the stratosphere. References Atkinson R.; Gas-phase tropospheric chemistry of organic compounds, J. Phys. Chem. Ref. Data, Monograph no.2 (1994). Kindler T.P., W.L. Chameides, P.H. Wine, D.M. Cunnold, F.N. Alyea and J.A. Franklin; The fate of atmospheric phosgene and the stratospheric chlorine loadings of its parent compounds: CCl4, C2Cl4, C2HCl3, CH3CCl3 and CHCl3, J. Geophys. Res. 100 (D1) (1995) 1235-1251. Nelson L., I. Shanahan, H.W. Sidebottom, J. Treacy and O.J. Nielsen; Kinetics and mechanism for the oxidation of 1,1,1-trichloroethane, Int. J. Chem. Kinet. 22 (1990) 577-590. Platz J., O.J. Nielsen, J. Sehested and T.J. Wallington; Atmospheric chemistry of 1,1,1-trichloroethane: UV absorption spectra and self-reaction kinetics of CCl3CH2 and CCl3CH2O2 radicals, kinetics of the reactions of the CCl3CH2O2 radicals with NO and NO2 and the fate of the alkoxy radical CCl3CH2O, J. Phys. Chem. 99 (17) (1995) 6570-6579. Sidebottom H., and J. Franklin; The atmospheric fate and impact of hydrofluorocarbons and chlorinated solvents, Pure and Appl. Chem. 68 (1996) 1757-1769. Talukdar R.K., A. Mellouki, J.B. Burkholder, M.K. Gilles, G. Le Bras and A.R. Ravishankara; Quantification of the tropospheric removal of chloral (CCl3CHO): Rate coefficient for the reaction with OH, UVabsorption and quantum yields, J. Phys. Chem. 105 (2001) 5188-5196. Tuazon E.C., R. Atkinson, S.M. Aschmann, M.A. Goodman and A.M. Winer; Atmospheric reactions of chloroethenes with the OH radical, Int. J. Chem. Kinet. 20 (1988) 241-265. Wenger J., S. Le Calvé, H.W. Sidebottom, K. Wirtz, M.M. Reviejo and J.A. Franklin; Photolysis of chloral under atmospheric conditions, Environ. Sci. Technol. 38 (2004) 831-837.

Simulation Chamber Study of the Oxidation of Acetic Acid by OH Radicals: Detection of Reaction Products by CW-CRDS in the Near-Infrared Range Sabine Crunaire1,3, Christa Fittschen1, Bernard Lemoine2, Alexandre Tomas3, and Patrice Coddeville3 1

PC2A, “Physico-Chimie des Processus de Combustion et de l’Atmosph è re”, UMR CNRS 8522 2 PhLAM, “Physique des Lasers, Atomes et Molécules”, UMR CNRS 8523 Université des Sciences et Technologies de Lille 1, 59655 V illeneuve d’A scq Cedex, France 3 Département Chimie-Environnement, École des Mines de Douai, 941 Rue Charles Bourseul, 59508 Douai Cedex, France Key Words: Acetic acid, CW-CRDS, OH radicals, Smog chamber

Abstract The branching ratio for the reaction of OH radicals with CH3COOD was determined at 298 r 2 K and at atmospheric pressure using an indoor smog chamber (300 L) coupled to two different detection systems: (i) Gas Chromatograph (GC) with FTIR detection for determination of [CH3COOD], (ii) Cavity Ring Down Spectroscopy (CRDS) using a telecommunication diode laser as a source for the determination of [HDO] and [H2O]. The reaction CH3COOD + OH occurs via D-atom abstraction with an efficiency of 36 r 20 %. Introduction Although anthropogenic emissions of volatile organic compounds (VOCs) in France seem to have been decreasing for several years, the contribution from oxygenated VOCs remains constant (CITEPA, 2000). Primary emissions of this type of compounds are mainly due to their use in industry (as paints, solvents, etc.). Some are also employed in fuels as substituents for aromatic hydrocarbons. In addition, as the atmospheric oxidation of hydrocarbons generally leads to oxygenated compounds (like carbonyls), secondary emissions are also significant. Among the oxygenated compounds, carboxylic acids have become the subject of growing interest over the past two decades. Low molecular weight carboxylic acids like acetic acid have been recognized as potentially important especially in urban polluted atmospheres where concentrations can exceed 20 µg/m3 (Chebbi and Carlier, 1996). Acetic acid is produced photochemically mainly from reactions of the peroxy acetyl radical (CH3CO3) with other peroxy radicals. For example, the reaction of CH3CO3 with HO2 is known to lead to about 20% CH3C(O)OH (Tyndall et al., 2001). Acetic acid in the gas phase is also produced by reaction of ozone with various olefins like propene, butene or pentene (Atkinson and Arey, 2003). A total photochemical source strength of 120 Tg/year has been reported (Baboukas et al., 2000). The contribution of direct emissions from anthropogenic (biomass combustion, motor exhaust) and biogenic (bacteria metabolisms, emission from soil and vegetation) sources is estimated at 48 Tg/year (Chebbi and Carlier, 1996). Acetic acid is principally removed from the atmosphere by wet deposition. In the atmospheric gas phase, the main loss of CH3C(O)OH is its reaction with OH radicals. 181 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 181–191. © 2006 Springer. Printed in the Netherlands.

S. Crunaire et al.

182 OH + CH3C(O)OH o products

(1)

The corresponding reaction rate, k1 = 7.5 x 10-13 cm3.molecule-1.s-1 at T = 298K (Singleton et al., 1989), corresponds to an atmospheric residence time (W k1[OH]) of more than a week (with typical tropospheric OH concentration of 1×106 molecule.cm-3). The photooxidation chain contributes to the production of photooxidants, so that the implications in the atmospheric HOx budget are important (Atkinson and Arey, 2003). However, models usually do not take these effects into account because too little is known about the fate of CH3C(O)OH in the atmosphere. In order to assess its tropospheric impact information on the product distribution of the reaction is needed. Two channels are expected to occur in the OH-initiated oxidation of acetic acid: OH + CH3C(O)OH o CH3 + CO2 + H2O o CH2C(O)OH + H2O

(1a) (1b)

According to Singleton et al. (1989) and Butkovskaya et al. (2004), hydrogen atom abstraction from the carboxyl group (pathway 1a) is the preferred pathway. A branching ratio of (64 r 17) % has been determined between 249 and 300 K (Butkovskaya et al., 2004). We present here the determination of the branching ratio of the products formed in the reaction of partly deuterated acetic acid with hydroxyl radicals under atmospheric conditions (298 K and 760 Torr) in a simulation smog chamber. OH + CH3C(O)OD o CH3 + CO2 + HDO o CH2C(O)OD + H2O

(2a) (2b)

The detection of CO2 was not possible at the time or writing due to a lack of an adequate light source (such studies are planned), however, the formation of HDO and H2O has been investigated. CH3COOH and CH3COOD have the same molecular structure, thus by analogy it is expected that the branching ratios will be fairly similar. The HDO (2a) and H2O (2b) formation rates were measuring by the CW-CRDS technique (Continuous Wave – Cavity Ring Down Spectroscopy). The setup was developed so that it could be coupled to the smog chamber. This is, to our knowledge, the first report of such a coupling and it will be described in detail in the next section. Experimental Section Simulation chamber (Figure 1) Experiments were performed in a Teflon-film bag of about 300 litres placed in a dark box equipped with an irradiation system and a temperature regulation device. The latter consist of 2 fans fixed at the top of the box flushing laboratory air around the bag. The irradiation system consists of 6 Philips TMX 204 LS fluorescent tubes (30 W, Omax = 254 nm) and 6 Philips TMX 200 LS tubes (18 W, Omax = 365 nm), distributed on both chamber sides. Each tube can be operated separately in order to modulate the irradiation flux. In the experiments presented below, only the fluorescent tubes at 365 nm have been used. Organic

Simulation Chamber Study of the Oxidation of Acetic Acid by OH Radicals

183

compounds are introduced into the smog chamber using a heated (T ~ 100°C) small glass chamber under vacuum. After vaporization, they are pressed into the bag by a flow of purified air (zero air). purified air

evacuation introduction of organic d

CW-CRDS GC-FTIR

heated glass chamber under vacuum

Teflon film bag

introduction of OH precursors

irradiation system: 6 fluorescent tubes at Omax = 254 nm and 6 tubes at Omax = 365 nm

Figure 1.

Top view of the indoor simulation chamber at the Ecole des Mines.

Chemicals For all the experiments, hydroxyl radicals were produced by the photolysis of methyl nitrite around 365 nm: CH3ONO + hQ (365 nm) o CH3O + NO CH3O + O2 o HO2 + HCHO HO2 + NO o NO2 + OH Methyl nitrite was prepared by slowly adding a dilute solution of H2SO4 to a mixture of NaNO2 and methanol according to the synthesis described by Taylor et al. (1980). Gas samples of about 20 mL of CH3ONO were injected in the chamber through a septum. Partly deuterated acetic acid CH3C(O)OD was obtained commercially (Acros Organics) with a purity higher than 98 % for the D-atom. Purified air was produced by flowing compressed air through a commercial air purification system (Claind AG 2301 HC). Experimental procedure Experiments were carried out at 298 r 2 K and at atmospheric pressure. Acetic acid was introduced in concentrations between 20 and 100 ppm and the bag was filled with purified air. Mixtures were left approximately one hour in the dark for stabilization during which time several samples were taken in order to determine accurately the initial concentration of reactant. The bag was then irradiated for one to four hours and sampling was

S. Crunaire et al.

184

performed at short and regular time intervals (between 15 and 30 minutes). After each run, the mixture was evacuated and the bag was then cleaned by filling it with purified air. This was repeated three times. Two different techniques were used to analyse the reaction mixture. The concentrations of CH3C(O)OD were determined using a GC-FTIR analytical device. The concentrations of reaction products HDO and H2O were measured by the CW-CRDS technique recently developed in our laboratory. Analysis apparatus 1: sample loop-TCT-GC-FTIR Sampling procedure and thermal desorption system (Figure 2) (a) Sample loop filling mode oven 150°C trap tube -190°C

(b) Trap mode oven trap tube

(c) Injection mode oven trap tube

150°C -190°C 30mL/min

50mL/min

150°C 280°C 30mL/min

sample loop

Teflon bag mixture

purified air

purified air

6 ports rotary valve

He

oven

liquid N2

He

He

empty glass tube "transfer line" cold trap tube filled with glass beads

vent

Figure 2.

GC 2 mL/min

30mL/min

GC 2 mL/min

GC 2 mL/min

Schematic diagrams of the device for sampling and thermal desorption.

Gas samples were taken from the bag using a 20 mL loop (4.57 mm i.d. Silcosteel tubing) and a 6-port valve (Valco). Figure 2a shows the loading position of the valve. The gas sample from the chamber is pumped through the heated (T = 100°C) sample-loop at 50 mL/min for 4 min. After switching off the pump, the 6-port valve is turned to the injection position (Figure 2b) and the 20 mL gas sample is conveyed to the cold trap of the TCT Chrompack (Thermal desorption and Cold Trap) device by the carrier gas. It should be noted

Simulation Chamber Study of the Oxidation of Acetic Acid by OH Radicals

185

that this system allows several fillings of the sample loop to increase the sampling volume. During the experiments, depending on the concentrations of reactants, injections of one or two loops have been carried out. The cryo-focusing with liquid nitrogen was enhanced for the more volatile compounds by filling the trap (uncoated deactivated fused silica; i.d. = 0.53 mm; length = 40 cm) with glass beads. Finally, Figure 2c shows the injection step. The trap tube was subject to a fast temperature increase (approximately 10°C/s) in order to inject the compounds into the gas chromatograph within a few seconds. GC-FTIR analysis The sample loop-TCT injection unit is coupled to a gas chromatograph (GC) Varian 3300 with detection by a FTIR spectrometer (Nicolet Magna 550) (Figure 3). The chromatographic separation was performed using a capillary column CP SIL 5 CB (50 m x 0.32 mm i.d.). The following temperature program was used for the GC oven: isothermal at 40°C for 5 min, first temperature ramped at 8°C/min to 120°C, second temperature ramped at 15°C/min to 225°C, where it was maintained for 5 min. Helium was used as the carrier gas at an initial pressure of 15 psi (40°C). The column effluent was passed through a fused silica transfer line (0.6m x 0.32mm i.d.) to the light pipe, both maintained at 240°C. The light pipe is a narrow borosilicate tube with a smooth, thin gold coating layer on the inside surface. By reflecting the IR beam through the light pipe via this coating, the path length of the cell is increased by a factor of ten or more with respect to the actual length of the tube (Ragunathan et al., 1999). The infrared beam was focused onto a fast response MCT detector cooled with liquid nitrogen.

TCT

Gas sample loop (20mL – T ~100°C)

Michelson interferometer

FID

capillary column

Teflon bag's sample

lightpipe

IR source

Principal infrared bench

Figure 3.

MCT detector

GC-FTIR interface

Transfer line

TCT-GC-FTIR analytical device.

GC

S. Crunaire et al.

186

Data acquisition was performed using compatible Omnic software (Nicolet). The Gram-Schmitt (total spectral response) chromatogram was constructed from interferograms registered at different times. Spectra were taken in real time between 650 and 4000 cm-1 at a resolution of 16 cm-1 with a scan every 2 s (average of 10 interferograms). Functional group chromatograms were extracted from the Gram-Schmitt chromatogram in the 1600-1900 cm-1 spectral region (C=O band) to carry out the quantitative analysis of acetic acid. Analysis apparatus 2: CW-CRDS Introduction Cavity Ring-Down Spectroscopy was introduced in 1988 by O'Keefe and Deacon as a spectroscopic method for absorption measurements (O'Keefe and Deacon, 1988). It is a versatile high sensitivity absorption technique. One of the most essential advantages of CRDS in contrast to usual absorption methods is that the CRDS signal is not affected by intensity fluctuations of the laser since only the decay time of the signal, which does not depend on the laser intensity, is detected. The setup of O'Keefe and Deacon employed a pulsed, tunable laser, and is known as a "generic pulsed CRDS setup". Since the publication of their article, the method of CRDS has been applied to different absorptions ranging from ~ 200 nm to ~ 10 µm and many variants of excitation and detection schemes have been used. Various methods and applications of CRDS are described in a number of reviews (Wheeler et al., 1998; Berden et al., 2000; Brown, 2003). In 1997, Romanini et al. adapted the use of small, continuous-wave diode lasers to CRDS (Romanini et al., 1997a; Romanini et al., 1997b). In order to perform quantitative analyses of H2O and HDO during oxidation of deuterated acetic acid in the smog chamber, we have developed a similar setup. CRDS principle The principle of CRDS is based upon a measurement of the rate of decay of an optical resonator with a high quality factor, indicating a long lifetime of the photons in the cavity. The principle is schematically illustrated in Figure 4. L

L

laser

laser

.. .

.. . M1

M2

(a) Figure 4.

time

M1

d

(b)

M2

time

Principle of CRD absorption measurements. (a) light decaying in an empty cavity, (b) cell is filled with an absorber.

After coupling the laser light into the cavity, the light intensity decreases with every pass within the resonator because of reflection losses of the mirrors and absorption losses of the medium. Therefore an exponential decay is obtained at the detector (eq.1).

Simulation Chamber Study of the Oxidation of Acetic Acid by OH Radicals § t· It I 0 ˜ exp¨  ¸ © W¹

187

(eq.1)

Without absorbing species between the mirrors (Figure 4a), the ring down time of the optically empty cavity is given by (eq.2): W0

L c ˜ (1  R )

(eq.2)

where c is the velocity of light and R the reflectivity of the mirrors. The decay time of the cavity filled with an absorbing medium (Figure 4b) is (eq.3): W abs

L c ˜ >1  R  Dd @

(eq.3)

where Dthe linear absorption coefficient is inversely proportional to the measured decay time of the CRDS signal (eq.4; here L=d).

D

1 § 1 1· ˜ ¨¨  ¸¸ c © W abs W 0 ¹

(eq.4)

Using D makes it possible to determine directly the concentration of a trace gas inside the cavity if the absorption cross-section Vabs is known (eq.5).

D

>abs@˜ V abs

(eq.5)

CW-CRDS setup and data acquisition

The experimental setup (Figure 5) used in this work was very similar to that described by Romanini et al. (1997b). It works with an external cavity diode laser (Sacher Lasertechnik) tunable from 1420 to 1480 nm with an output power of at least 3 mW. For the cavity mirrors, a 1 m radius of curvature and low-loss mirrors by Layertec, covering the range from about 1400 nm to 1550 nm, were used. They were placed at a distance L of 70 cm and sealed the cell. One of them was mounted on a piezoelectric transducer (Physik Instrumente). This device was used to modulate the cavity length in order to let one of the cavity modes oscillate around the laser line. As the beam reflected by the input cavity mirror can strongly perturb the laser, a Faraday isolator (Isowave) was placed in the path to attenuate this effect. The intensity of the light leaving the cavity was measured with a photodiode avalanche. When a predefined threshold was reached, an acoustic-optical modulator (Gooch and Housego) turned off the laser beam and a clean ring-down decay could be measured. The ring-down signals were registered by an InGaAs photodiode (EG&G) and digitised by a 12 bits PC interface card (National Instruments). Using the graphic programming language LabVIEW (National Instruments), the ring-down signals were recorded at a rate of about 100 Hz. The analysis of each ring-down (fitted by an exponential function using a nonlinear Levenberg-Marquardt algorithm) is made separately before averaging the decay time W. With the system described above, a section of the C2H2 overtone spectrum around 1512 nm has been measured, using a distributed feedback diode laser. The noise equivalent absorption coefficient is in the order of 5×10-8 cm-1 or a detection limit of less than 100 ppb

188

S. Crunaire et al.

for C2H2 under the experimental conditions. A summary of the performance characteristics of the setup is given in Table 1.

1 meter

Wavemeter

1468.584 nm

0.75 meter

Fiber optic optic Fiber L

ECDL

AOS

OI

Trigger system Gate generator

1st order

L

L

L PZT

Glass cell APD T, P L

MC

M CC

P

PC

N2 or He

Threshold detector

gas or or smog smog Standard gas chamber's sample sample chamber's

A/D Converter working with LabVIEW 6.1TM

ECDL: External Cavity Diode Laser; OI: Optical Isolator; L: Lens; AOS: Acousto-Optic Switch; PZT: Piezo-Electric actuator; MC: Cavity mirror; APD: Avalanche Photodiode; PC: Pressure Controller; P: Pump

Figure 5.

Scheme of the CW-CRDS setup

Table 1.

Summary of the performance characteristics of the CW-CRDS setup.

coarse tunability of the ECDL fine tunability of the ECDL cavity length cavity mirrors

between 1420 and 1490 nm on 1 nm using the grating piezo control input 70 cm with actual cell RC = 1 m; R ~ 99.99 % o W0 = 25 µs

~ 7 km without absorbing species frequency of cavity length > 200 Hz o one ringmodulation down event every 2.5 ms spectral resolution > 0.01 cm-1 effective path length

sensitivity of the system

5.10-8 cm-1

Determination of [HDO] and [H2O] during smog chamber experiments For quantitative determination of the HDO and H2O concentrations before and during oxidation of CH3COOD by OH in the smog chamber, the reaction mixture was pumped with a flow rate of 70 mL/min through the cavity. During a measurement, the cavity pressure is held constant at 50 Torr to reduce the broadening of the two spectral lines and thus increase the selectivity. Figure 6 shows a CW-CRDS spectrum of characteristic spectral lines of H2O and HDO near 6776 cm-1 (without any averaging). From this measurement, the corresponding concentrations are calculated using the database HITRAN 2004 (Rothman et al., 2005). It takes approximately one minute to measure a spectrum as presented in Figure 6, however, for real-time detection it is sufficient to measure the absorption at only a few wavelengths. Thus

Simulation Chamber Study of the Oxidation of Acetic Acid by OH Radicals:

189

the concentrations can be determined within a few seconds. Note that quantitative analysis does not require any calibration, CW-CRDS leads directly to absolute concentrations, if the absorption cross section of the measured spectral line is known (from HITRAN 2004 or GEISA databases for example).

+ experimental points Gaussian fit

HDO

H2 O

Figure 6.

CW-CRDS spectrum obtained at 50 Torr from a smog chamber mixture.

Preliminary Results and Discussion The study of the reaction OH + CH3COOD was conducted in an excess of CH3ONO over CH3COOD. The branching ratio of channel (2a) was determined as the ratio of the concentration of HDO produced to the concentration of CH3COOD consumed in reaction 2: R

k 2a k 2a  k 2b

k 2a k2

d [ HDO ]  d [ CH 3 COOD ]

>HDO @formed [ CH 3 COOD ] reacted

Figure 7 shows an example of a plot of the formation of HDO and H2O versus the loss of CH3COOD for one experiment. The accuracy of the measured CH3COOD concentration was determined from the statistic error (2V) of the calibration curve, obtained by liquid injections. It was estimated to be 5 %. Uncertainties in the HDO and H2O concentrations arise mainly from errors in the measured absorbance and from the accuracy in the absorptions cross sections. They were estimated to be 10 %. The slopes of the above lines give directly the branching ratios R2a and R2b. At first, the branching fraction R2a = (36 r 20) % indicates that channel 2a would be the minor oxidation channel for the OH + CH3COOD reaction. However, this result can't be easily compared with the one of Butkovskaya et al. (2004) because too little is known about the

190

S. Crunaire et al.

isotope effect in the reaction of CH3COOD and OH. On the one hand, the yield of formation of H2O is too high (R2b > 1) and other sources of production must be taken into account. Principally, the interactions between H2O molecules and the walls of the chamber can be responsible for a large perturbation in the H2O absorption signal. Furthermore, the reactions of OH with other species like CH3ONO, HCHO (product of CH3ONO photolysis) or CH3OH (impurity in the CH3ONO synthesized) can be supplementary sources of H2O.

3 '[HDO] and '[H2O] (molecules/cm )

4,5E+15 4E+15

HDO

3,5E+15

H2O

y = 2,2763x R2 = 0,983

3E+15 2,5E+15 2E+15 1,5E+15

y = 0,3627x

1E+15

R2 = 0,6895

5E+14 0 0

2E+14

4E+14

6E+14

8E+14

1E+15

1,2E+15 1,4E+15 1,6E+15 1,8E+15

-'[CH3COOH] (molecules/cm3)

Figure 7.

Formation of HDO (crosses) and H2O (open circles) versus loss of CH3COOD following irradiation of a CH3COOD (~ 3,5×1015 molecules/cm3) + CH3ONO (~ 4×1018 molecules/cm3) mixture with zero air as diluent at 298 r 2 K and atmospheric pressure. The solid lines are linear regressions of the two data sets.

Conclusions The present study reports the first coupling between a smog chamber and a CW-CRDS setup. After a successful evaluation the CW-CRDS setup in the laboratory (spectroscopic study of C2H2 around 1512 nm), this new analytical device permits us to follow the production of small molecules (CO, CO2, C2H2, HCOOH, etc.) and also of isotopic compounds (HDO, H218O for example) almost in "real time" during experiments on VOC degradation (oxidation or photolysis) in a simulation chamber. For the first experiments carried out on the oxidation of CH3COOD by OH radicals, a branching ratio of (36 r 20) % has been determined for the D-atom abstraction channel. These experiments have revealed several difficulties in the analysis of the results and comparison with literature data because of the unknown reaction mechanisms and kinetics involving deuterated compounds. Additional experiments are necessary to resolve these difficulties. Determination of the branching ratio for the product channels in reaction of CH3COOH with OH radicals in a smog chamber under atmospheric conditions by measuring the CO2 formation yield (CW-CRDS using a distributed feedback diode laser at 1570 nm) is in progress.

Simulation Chamber Study of the Oxidation of Acetic Acid by OH Radicals:

191

Acknowledgments The authors gratefully acknowledge the help of T. Leonardis (Ecole des Mines – Douai) and Dr. D. Romanini (Laboratoire de Spectrométrie Physique – Grenoble) for their technical assistance. Moreover, this work has been achieved with the supports of the CPER (Contrat de Plan Etat Région) Nord-Pas de Calais "air quality", the CNRS (Centre National de la Recherche Scientifique), the CERLA (Centre d'Etudes et de Recherches Lasers et Applications) and the FEDER (Fonds Européen de DEveloppement Régional). References Atkinson R. and J. Arey; Atmospheric degradation of volatile organic compounds, Chem. Rev. 103 (2003) 46054638. Baboukas E. D., M. Kanakidou and N. Mihalopoulos; Carboxylic acids in gas and particulate phase above the Atlantic Ocean, J. Geophys. Res. 105 (2000) 14459-14471. Berden G., R. Peeters and G. Meijer; Cavity ring-down spectroscopy: Experimental schemes and applications, Int. Rev. in Phys. Chem. 19 (2000) 565-607. Brown S. S.; Absorption spectroscopy in high-finesse cavities for atmospheric studies, Chem. Rev. 103 (2003) 5219-5238. Butkovskaya N. I., A. Kukui and N. Pouvesle, G. Le Bras; Rate constant and mechanism of the reaction of OH radicals with acetic acid in the temperature range of 229-300 K, J. Phys. Chem. A 108 (2004) 7021-7026. Chebbi A. and P. Carlier; Carboxylic acids in the troposphere, occurrence, sources, and sinks: a review, Atmos. Environ. 30 (1996) 4233-4249. CITEPA : Emissions dans l’air en France, comparaisons de données d'émissions de la France et d'autres pays, URI: 'http://www.citepa.org/emissions/france_autres/Comparaisons.pdf' (2000). HITRAN 2004; URI: 'http://www.hitran.com'. O'Keefe A. and D. A. G. Deacon; Cavity ring-down optical spectrometer for absorption measurements using pulsed laser sources, Rev. Sci. Instrum. 59 (1988) 5244-2551. Ragunathan N., K. A. Krock, C. Klawun, T. A. Sasaki and C. L. Wilkins; Gas chromatography with spectroscopic detectors, J. Chromatogr. A 856 (1999) 349-397. Romanini D., A. A. Kachanov, N. Sadeghi and F. Stoeckel; CW cavity ring down spectroscopy, Chem. Phys. Lett. 264 (1997a) 316-322. Romanini D., A. A. Kachanov and F. Stoeckel; Diode laser cavity ring down spectroscopy, Chem. Phys. Lett. 270 (1997b) 538-545. Rothman L.S., D. Jacquemart, A. Barbe, D. Chris Benner, M. Birk, L.R. Brown, M.R. Carleer, C. Chackerian Jr., K. Chance, V. Dana, V.M. Devi, J.-M. Flaud, R.R. Gamache, A. Goldman, J.-M. Hartmann, K.W. Jucks, A.G. Maki, J.-Y. Mandin, S.T. Massie, J. Orphal, A. Perrin, C.P. Rinsland, M.A.H. Smith, J. Tennyson, R.N. Tolchenov, R.A. Toth, J. Vander Auwera, P. Varanasi, and G. Wagner; The HITRAN 2004 Molecular Spectroscopic Database, JQSRT, Preprint version, 2005. Singleton D. L., G. Paraskevopoulos and R. S. Irwin; Rates and mechanism of the reactions of hydroxyl radicals with acetic, deuterated acetic, and propionic acids in the gas phase, J. Am. Chem. Soc. 111 (1989) 52485251. Taylor W. D., T. D. Allston, M. J. Moscato, G. B. Fazekas, R. Kozlowski and G. A. Takacs; Atmospheric photodissociation lifetimes for nitromethane, methyl nitrite and methyl nitrate, Int. J. Chem. Kinet. 12 (1980) 231-240. Tyndall G. S., R. A. Cox, C. Granier, R. Lescaux, G. K. Moortgat, M. J. Pilling, A. R. Ravishankara and T. J. Wallington; Atmospheric chemistry of small organic peroxy radicals, J. Geophys. Research. 106 (2001) 12157-12182. Wheeler M. D., S. M. Newman, A. J. Orr-Ewing and M. N. R. Ashfold; Cavity ring-down spectroscopy, J. Chem. Soc., Faraday Trans. 94 (1998) 337-351.

Kinetics, Products and Mechanism of O(3P) Atom Reactions with Alkyl Iodides Ian Barnes Bergische Universität Wuppertal, Fachbereich C - Physikalische Chemie, Gauß Straße 20, D-42119 Wuppertal, Germany

Key Words: Alkyl iodides, Halogens, Oxygen atoms, Relative kinetic, Reaction mechanisms

Introduction Alkyl halides (RX: X = Cl, I) are an important source of halogens in the atmosphere. The major tropospheric sinks of these compounds are photolysis (RBr, RI) and reaction with OH radicals. In the case of alkyl iodides (RI) relative kinetic studies of their OH reactions in photoreactors are complicated by fast reactions with the O(3P) atoms generated by the photochemical OH radical sources. Figure 1 below shows a ln-ln plot of the kinetic data from an experiment performed in a large photoreactor to determine the OH rate coefficient for the reaction OH + CH3CH2CH2I relative to OH + ethene using the photolysis of methyl nitrite (CH3ONO) as the OH radical source. A recent example of the implementation of the relative kinetic technique for the determination of OH radical rate coefficients in a photoreactor can be found in Olariu et al. (2000).

ln((Io/I) propyl iodide

0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 0.00

0.10

0.20

0.30

0.40

0.50

0.60

ln(Io/I) ethene

Figure 1.

Plot of the kinetic data obtained from the irradiation of a propyliodide, ethene, CH3ONO, and synthetic air reaction mixture in a 420 l photoreactor.

The non-linearity with increasing reaction time in the plot in Figure 1 is due to very fast reactions of RI with O(3P). An analysis of the initial slope yields a rate coefficient for the reaction of OH with 1-propyliodide which is in reasonable agreement with the recommended literature value for the reaction of OH with 1-propyliodide (Carl and Crowley, 2001; Cotter et al., 2003) while the slope at extended reaction time yields a rate coefficient in reasonable agreement with literature values for the reaction of O(3P) with 1-propyliodide (Gilles et al., 1996; Terual et al., 2004). Luo et al. (1995) have successfully employed the relative kinetic technique in a 500 l Teflon chamber to measure rate coefficients for the reactions of O(3P) with a series of alkenes 193 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 193–205. © 2006 Springer. Printed in the Netherlands.

194

I. Barnes

and terpenes at ambient temperature and atmospheric pressure. The experiments were performed in N2 bath gas and an excess of both NO2 and NO to optimise the concentration of O(3P) and depress possible interferences by reactions of OH and NO3 radicals and O3 with the compounds under investigation. The few rate coefficients for the reactions of O(3P) reactions with alkyl iodides, which have been measured, have been found to be extremely rapid (Gilles et al., 1996; Terual et al., 2004; and references therein). The measured rate coefficients are much faster than those for O(3P) with the chlorine and bromine analogs and are also much faster than those for OH radicals with the alkyl iodides (Carl and Crowley, 2001; Cotter et al., 2001). This has been explained by invoking the initial formation of a R–I–O complex. IO elimination is the most important product channel for the small iodoalkanes such as CH3I and CF3I, however, for longer chain iodoalkanes the presence of a C–H bond in the E-position to the I-atom enables the intramolecular abstraction of an H-atom via a five-membered ring transition state to eliminate HOI and form an alkene (Klaassen et al., 1996; Loomis et al., 1997). As discussed above the kinetic database for the reactions of O(3P) with alkyl iodides is relatively limited. The higher reactivity of alkyl iodides to O(3P) atoms compared to OH radicals suggests that the application of the relative kinetic technique in a photoreactor, as applied by Luo et al. (1996) for the determination of O + alkene/terpene rate coefficients, should also be applicable for the determination of rate coefficients for O + alkyl iodides. Because of the much higher rate coefficients for O(3P) with the alkyl iodides compared to OH, inference by any OH radicals produced in the system will be minimal. Reactions with O3 and NO3 radicals are too slow to be of any importance. In the present work the relative kinetic technique has been applied in large and small photoreactors to measure rate coefficients for the reaction of O(3P) atoms with a series of alkyl iodides at room temperature and atmospheric pressure. The products formed in N2 have also been investigated. Experimental Product and kinetic experiments have been performed on the reaction of O(3P) with a series of mono- and disubstituted alkyl halides. The compounds investigated are listed in Table 1 in the Results section. The product studies were all carried out in either a 1080 1 quartz glass chamber (Mihalopoulos et al., 1992) or a 480 l cylindrical Duran glass chamber (Barnes et al., 1993). Schematic outlines of the 1080 l
and 480 l chambers are shown in Figures 2 and 3, respectively. In the 1080 l
chamber an internally mounted White type multiple reflection mirror system operated at a total optical path length of (484.7 r 0.8) m coupled to an FTIR spectrometer was used for reactant and reference monitoring. IR spectra were recorded at a spectral resolution of 1 cm-1 using a Nicolet NEXUS FT-IR spectrometer equipped with a MCT detector. The 480 l chamber is also equipped with 3 built-in White mirror systems for long path (51.6 m) FTIR (Nicolet Magna 520, MCT detector, 1 cm-1 resolution), VIS (28.8 m) and UV (28.0 m) in situ monitoring of reactants and products. The majority of the kinetic experiments have been carried out in a simple 20 l reactor using GC-FID for the analysis; Hewlett Packard 5890 Series II with HP 3395 integrator, separation on a RESTEX RTX-1301 column (30 m long, 0.53 mm inner diameter and 0.25 Pm film thickness). This reactor is shown in Figure 4. Samples were taken using a 1 ml gas-

Kinetics, Products and Mechanism of O( 3P) Atom Reactions with Alkyl Iodides

195

tight syringe and injected into the GC system. For some of the larger alkyl iodides control experiments were also performed in the 480 l chamber using FT-IR absorption spectroscopy for the kinetic analysis.

Figure 2.

Schematic outline of a 1080 l quartz glass photoreactor equipped with an internally mounted White type mirror system for FTIR in situ monitoring of reactants and products. The chamber can be temperature regulated in the range 303 to 283 K. The chamber is surrounded by 32 super actinic fluorescent lamps (Philips TL 05/40 W: 320 < O < 480, Omax = 360 nm) and 32 low pressure mercury lamps (Philips TUV 40 W: Omax = 254 nm) mounted evenly around the reactor. Other types of lamps can be easily mounted.

All experiments were performed at 1013 mbar total pressure of N2 at 298 ± 2 K using the photolysis of NO2 with fluorescent lamps (Philips TL 40W/05, 320 < O < 450 nm) as the O(3P) atom source: NO2 + hv • NO + O Rate coefficients for the reaction of O(3P) with selected alkyl iodides were determined using the relative kinetic technique as outlined by Luo et al. (1996). 2-Iodopropane (CH3CH(I)CH3) was used as the reference compound: O + RI o products; k1 O + 2-iodopropane o products; k2

I. Barnes

196

Providing that RI and 2-iodopropane are removed solely by reaction with O atoms the following rate law holds: ­[ RI]t 0 ½ ln ® ¾ ¯ [RI]t ¿

k1 ­[2 - iodopropane ]t 0 ½ ln ® ¾ k 2 ¯ [2 - iodopropane ]t ¿

(eq 1)

where [RI]to and [2-iodopropane]to are the concentrations of the alkyl iodide under test and reference alkyl iodide, respectively, at time to, [RI]t and [2-iodopropane]t are the corresponding concentrations at time t, and k1 and k2 are the rate coefficients of the alkyl iodide and reference compound with O(3P), respectively. Plots of ln([RI]to/[RI]t) against ln([2iodopropane]to/[2-iodopropane]t) should straight lines with a slope of k1/k2 and zero intercept. The rate coefficients for the reactions of O(3P) with the various alkyl iodides were placed on an absolute basis using a rate coefficient value of k = 4.97 × 10-11 cm3 molecule-1 s-1 from Teruel et al. (2004) for the reaction of O(3P) with 2-iodopropane. This value is in good agreement with the other reported value by Gilles et al. (1996).

Figure 3.

Schematic outline of a 480 Aquartz glass photoreactor equipped with an internally mounted White type mirror systems for FTIR, VIS and UV in situ monitoring of reactants and products. The chamber is surrounded by 20 super actinic fluorescent lamps (Philips TL 05/40 W: 320 < O < 480, Omax = 360 nm) and 12 gold fluorescent lamps (Philips TLD 16/36 W: 500 < O < 700 nm).

The approximate initial reactant concentrations in both reactors (in ppmv units) were as follows: NO2, 50-100; alkyl iodide 10-30 and 2-iodopropane 10-30 (1 ppmV = 2.46 u 1014 molecule cm-3 at 298 K and atmospheric pressure). The concentration-time behaviour of the alkyl iodides were monitored over a 15-30 min irradiation period in both the kinetic and product studies.

Kinetics, Products and Mechanism of O( 3P) Atom Reactions with Alkyl Iodides

197

Figure 4. Schematic outline of a 20 l glass reactor with a metal top flange for inlet and outlet ports. The reactor is equipped with 4 fluorescent lamps (Philips TL05/40 W) and can be evaluated to 0.01 mbar with a simple rotary pump.

Figure 5. Photograph of the simple 20 l glass reactor with GC analysis system shown schematically in figure 4.

Figure 5. Photograph of the simple 20 l glass reactor with a GC analysis system shown schematically in Figure 4.

Results and Discussion Kinetics O + RI To test the relative kinetic method in the small 20 l reactor the rate constant ratio for the reaction of O(3P) with 1-iodopropane relative to 2-iodopropane was determined. The kinetic data from a typical experiment plotted according to equation (1) are shown in Figure 6. A good linear correlation with zero intercept was obtained. The value of the rate coefficient obtained for the reaction of O(3P) with 1-iodopropane is in excellent agreement with the absolute value obtained by Teruel et al. (2004) using pulsed laser photolysis-resonance Fluorescence (PLP-RF) which demonstrates the validity of the applied method for the determination of rate coefficients for the reactions of O(3P) with alkyl iodides.

I. Barnes

198 0.70

Figure 6. Plot of the kinetic data according to eq. (1) for measurements on the reaction of O(3P) with 1iodopropane relative to that with 2-iodopropane.

ln(Io/I) 1-iodopropane

0.60 0.50 0.40 0.30 0.20 0.10 0.00 0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

ln(Io/I) 2-iodopropane

Examples of the kinetic data for some of the alkyl iodides investigated plotted according to eq. (1) are shown in Figure 7. In the main good straight line correlations were obtained with zero intercepts. The rate coefficients obtained for the alkyl iodides investigated are listed in Table 1 where they are compared with the available literature data. The rate coefficients determined in this work are estimated to have a total error of ± 15% 3.50

ln(Io/I) alkyl iodide

3.00 2.50 2.00

iodocyclopentane 1,4-diiodobutane 2-iodo-2-methylpropane 1-iodo-2-methylpropane 1,3-diiodopropane iodocyclohexane Linear (1,4-diiodobutane) Linear (iodocyclopentane) Linear (2-iodo-2-methylpropane) Linear (1-iodo-2-methylpropane) Linear (iodocyclohexane) Linear (1,3-diiodopropane)

Figure 7. Examples of plots of the kinetic data according to eq (1) for measurements on the reaction of O(3P) with several alkyl iodides relative to that with 2iodopropane.

1.50 1.00 0.50 0.00 0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

ln(Io/) 2-iodopropane

Rate coefficients have been determined for the reaction of O(3P) with 17 alkyl iodides 12 of which are first-time measurements. As can be seen from Table 1 there is good agreement between the rate coefficients reported in the literature for the C1 to C4 alkyl monoiodides (Gilles et al., 1996; Teruel et al., 2004) and those determined in this work. This good agreement adds further weight to the soundness of the applied experimental approach. The present work has considerably expanded the number and type of alkyl iodides investigated. The reactivity trends for the reactions of O(3P) with RI have been discussed in detail by Teruel et al. (2004) and Gilles et al. (1996). Teruel et al. (2004) have also made comparisons with the analogous reactions with OH radicals and Cl atoms. Therefore, only salient features of the reactivity trends and supplementary information will be presented here.

Kinetics, Products and Mechanism of O( 3P) Atom Reactions with Alkyl Iodides

199

As in the other studies a gradual increase in the rate coefficient is observed in progressing from CH3I to CH3CH2I to CH3CH2CH2I. It has been observed in this study that introducing a fourth carbon to the linear chain results in a further small increase in the rate coefficient, however, extensions of the chain to C5 and C6 have no further influence on the rate coefficient. In agreement with Teruel et al. (2004) it is found that the rate coefficient for the reaction of O(3P) with 2-iodopropane is greater than that for 1-iodopropane. It has also been found in the present work that the rate coefficient for the reaction of O(3P) with 2-iodobutane is higher than that for 1-butane. This behaviour is reconcilable in terms of the proposed main reaction mechanism mentioned in the Introduction, i.e. the secondary iodides have a larger number of E H-atoms that can be internally abstracted via a five-membered ring R-I-O complex transition state. For the C5 compounds 1-iodo-2-methylbutane and 1-iodopentane the rate coefficient for O(3P) with 1-iodo-2-methylbutane is higher than that for 1-iodopentane, presumably for similar reasons. The highest rate coefficient has been found for the reaction of O(3P) with 2-iodo-2-methylpropane which has the highest possible number abstractable E Hatoms. A number of di-iodo compounds have been investigated. In general the rate coefficients for the di-iodo compounds with chain lengths of •C2 are about a factor of 2 higher than the corresponding mono-iodo compounds. In the case of CH3I and CH2I2 the difference reported by Teruel et al. (2004) for the rate coefficients is nearly 4. The rate coefficient for the reaction of O(3P) with neopentyl iodide ((CH3)3CCH2I) is slightly higher than that for CH3I. Neopentyl iodide does not possess any E H-atoms and the mechanism is, as in the case of CH3I, an overall I abstraction mechanism to form IO. This conclusion is supported by the observations in the product studies. The rate coefficients for the reaction of cyclopentane and cyclohexane with O(3P) are larger than those for the corresponding linear chain 1-iodo-alkanes, which probably reflects the lower C-H bond strength in the cyclic compounds compared to the straight chain analogues. The measured rate coefficient for the reaction of O(3P) with cyclopentane in this work is, however, around 30% higher than that for cyclohexane despite the C-H bond energy in cyclohexane (399.6 kJ mol-1) being somewhat lower than that in cyclopentane (403.5 kJ mol-1). Stereo factors may possibly play a role in determining the rate of the reaction; cyclohexane has a chair conformation with zero ring strain whereas cyclopentane has an envelope conformation. This is, however, presently only conjecture, independent conformation of the rate coefficients for the two substances by another method is required before any positive conclusions can be drawn. The room temperature rate coefficients for O(3P) with fluorinated alkyl iodides (Gilles et al., 1996) and selected alkyl iodides (Teruel et al., 2004) and have found to show a linear correlation with the ionisation potential (IP) of the compounds. Table 2 lists the ionization potentials (IP) for a series of alkyl iodides taken from the NIST Chemistry WebBook (http://webbook.nist.gov/chemistry/) and their corresponding rate coefficients for reaction with O(3P) as measured in this study and taken from other available literature sources. The natural logarithm of the 298 K rate coefficients for– the compounds listed in Table 2 are plotted as a function of ionisation potential in Figure 8. The numbers on the plot refer to the correspondingly numbered compounds in Table 2. The points for compounds 3 and 9 overlap on the diagram.

I. Barnes

200 Table 1.

Summary of the product and kinetic data from studies on the reactions of O(3P) with alkyl halides at 1000 mbar total pressure of N2 and 298 ± 2K. The rate coefficients determined in this work (column (a)) have an estimated overall uncertainty of ± 15%. The alkene yields have an uncertainty of approximately ± 20%. k(O + RI) x 1011

Compound

Products

(cm3 molecule-1 s-1)

(a)

(b)

(c)

Alkenes (d) / (% mol yield)

Alkanes/other

iodomethane

CH3I

1.82 2.01

-

-

ethane

iodoethane

CH3CH2I

3.19 3.51

-

ethene / 100

-

1,2-diiodoethane

ICH2CH2I

6.25

-

Iodoethane / (f)

?

-

ethene / 42 1-iodopropane

CH3CH2CH2I

3.88 3.79

-

2-iodopropane

(CH3)2CHI

ref

1,3-diodopropane

ICH2CH2CH2I

11.5

-

-

allyliodide / 100

allyliodide

CH2=CHCH2I

10.2

-

-

1,5-hexadiene / (f)

3-iodopropylene oxide

1-iodobutane

CH3CH2CH2CH2I

4.83

-

-

1-butene / 70

n-octane

2-iodobutane

CH3CH2CHICH3

6.64

-

-

1-butene / 46

3,4-dimethylhexane

4.97 5.18

propene / 80

n-hexane

propene / 68

2,3-dimethylbutane

trans-2-butene / <10

1-iodo-2methylpropane

(CH3)2CHCH2I

4.39

-

4.54

isobutene / 65

2,5-dimethylhexane

2-iodo-2methylpropane

(CH3)2C(I)CH3

14.3

-

-

isobutene / 65

2,2,3,3-tetramethylbutane

8.42

-

-

4-iodo-1-butene / (f)

4-iodo-1,2-butylene oxide

1,4-diiodobutane ICH2CH2CH2CH2I

butadiene / 55(e)

1-iodopentane

CH3(CH2)4I

4.50

1-iodo-2methybutane

CH3CH2CH(CH3)CH2I 5.36

neopentyliodide

(CH3)3CCH2I

-

-

1-pentene / (f)

n-decane

2-methyl-1-butene / 100

3,6-dimethyloctane

2.50

-

-

(g)

2,2,5,5-tetramethylhexane

iodocyclopentane cyclo-C5H9I

9.27

-

-

cyclopentene/ (f)

bicyclopentyl

1-iodohexane

CH3(CH2)5I

4.75

-

-

1-hexene/ (f)

n-dodecane

iodocyclohexane

cyclo-C6H11I

6.91

-

-

cyclohexene / 68%

(a) This work; the values were put on an absolute basis using k = 4.97 x 10-11 cm3 molecule-1 s-1 for O(3P) with 2-iodopropane (ref) from Teruel et al. (2004); (b) Values from Teruel et al. (2004); (c) Values from Gilles et al. (1996); (d) Yield corrected for secondary reaction with O atoms; (e) Yield increased with time; the given yield is the yield at end of the experiment; (f) Observed but yield not determined; (g) No evidence for the formation of an alkene.

Kinetics, Products and Mechanism of O( 3P) Atom Reactions with Alkyl Iodides

201

The points in Figure 8 taken collectively indicate a distinct curvature in the relationship. However, if the points for the fluorinated compounds 14 and 16 are removed a reasonably linear correlation between the points for the remaining compounds is obtained.

-1 -1

rate coefficient (cm molecule s )

1.00E-09

4

1.00E-10 6

7

3

2

5 10 9 3

8 12 11

1.00E-11

13

14

16

15

1.00E-12 8.8

9

9.2

9.4

9.6

9.8

10

10.2

10.4

10.6

10.8

ionisation potential (eV)

Figure 8.

Plot of the rate coefficients for the reactions of O(3P) with a series of alkyl iodides as a function of the ionisation potential of the alkyl iodide. The numbers on the plot refer to the correspondingly numbered compounds in Table 2.

Products O + RI With the exception of iodomethane (CH3I) and neopentyl iodide ((CH3)3CCH2I) all the reactions of O(3P) with the alkyl iodides were found to produce as major product an alkene. An example is shown for 1-iodobutane in Figure 9. Hypoiodous acid (HOI), the probable coproduct in each case, was also identified via its known infrared absorptions at 3620 and 1070 cm-1 (Barnes et al., 1992). This compound is, however, very short-lived under the conditions of the experiments and was only visible in the infrared spectrum in the initial stages of the reaction (Figure 10) and could therefore not be quantified. For many of the alkenes calibrated reference spectra were available and the yield of the alkene has been determined. Oxygen atoms also react fairly rapidly with alkenes, therefore, the reaction of O atoms with the alkenes has been taken into consideration when calculating the alkene yields from the O + RI reactions. The yields and the identity of the alkene for the respective O + RI reactions are listed in Table 1.

I. Barnes

202 Table 2.

Comparison of the ionization potentials (IP) of selected alkyl iodides with the measured rate coefficients for the compounds with O(3P) atoms. The IPs’ are taken from the NIST Chemistry WebBook (http://webbook.nist.gov/chemistry/).

No Compound

Formula

1

neopentyliodide

(CH3)3CCH2I

2

1-iodohexane

CH3CH2CH2CH2CH2CH2I

3

1-iodopentane

4

Ionization Potential (eV)

rate coefficient k(O3P) 3 (cm molecule-1 s-1)

not known

2.60 · 10-11

9.179

4.75 · 10-11

CH3CH2CH2CH2CH2I

9.2

4.50 · 10-11

2-iodo-2-methylpropane

(CH3)2C(I)CH3

8.98

1.49 · 10-10

5

1-iodobutane

CH3CH2CH2CH2I

9.23

5.06 · 10-11

6

2-iodobutane

CH3CH2CHICH

9.1

6.84 · 10-11

7

2-iodopropane

(CH3)2CHI

9.175

4.97 · 10-11

8

1-iodopropane

CH3CH2CH2I

9.26

3.79 · 10-11

9

1-iodo-3-methylpropane

(CH3)2CHCH2I

9.202

4.54 · 10-11

10 1-iodoethane

CH3CH2I

9.35

3.50 · 10-11

11 iodomethane

CH3I

9.538

1.74 · 10-11

12 d3- iodomethane

CD3I

9.538

2.01 · 10-11

13 1,1,1-trifluoro-2-iodoethane

CF3CH2I

9.998

1.04 · 10-11

14 1,1,1-trifluoro-2-fluoro-2iodoethane

CF3CHFI

10.32

6.05 · 10-12

15 trifluroiodomethane

CF3I

10.23

4.34 · 10-12

16 pentafluoroethyliodide

CF3CF2I

10.66

4.82 · 10-12

In cases where calibrated alkene spectra were not available identification of the alkene was based on the infrared spectra taken from the NIST infrared database. With the exceptions of iodomethane (CH3I) and neopentyl iodide ((CH3)3CCH2I) the yield of the alkene from the various alkyl iodides was substantial in every case and for some compounds approached, within the error limits, unit yield, e.g. iodoethane, 1,3-diodopropane and 1-iodo-2methylbutane. In every reaction investigated absorptions in the ir spectra were observed which could be attributed to formation of an alkane formed from the recombination of the alkyl radicals produced via abstraction of the I atom from the RI compound, i.e. RI + O o R + IO R + R + N2 o R-R + N2

Kinetics, Products and Mechanism of O( 3P) Atom Reactions with Alkyl Iodides

203

0.50

absorbance (base 10)

0.40

0.30

d) c)

0.20

b) 0.10

a) 0.00 1300

1250

1200

1150

1100

1050

1000

950

900

850

800

750

700

-1

wavenumber (cm )

Figure 9.

Plot of the spectral data in the range 1300 – 700 cm-1 obtained in the irradiation of a 1-iodobutane (CH3CH2CH2CH2I)/NO2/N2 reaction mixture: a) spectrum of initial mixture before irradiation showing absorption bands from 1-iodobutane, b) the product spectrum after 10 minutes irradiation, c) difference spectrum of b) - a) with absorptions of 1-iodobutane zeroed, and d) reference spectrum of 1butene.

0.02

Intermediate formation of HOI

0.01

absorbance (base 10)

0.01 0.00 -0.01 -0.01 -0.02 -0.02

Decay of 1-iodobutane

-0.03 -0.03 1250

1150

1050

950

850

-1

wavenumber (cm )

Figure 10.

A difference spectra formed from the subtraction of a spectrum of a mixture of 1-iodobutane/NO2/N2 recorded before irradiation from one recorded after 4 minutes irradiation. The weak absorption with its centre at 1070 cm-1 is attributable to hypoiodous acid.

I. Barnes

204

The yields of the alkane products have not been quantified. The yield of the alkane will depend upon the experimental conditions used since the alkyl radicals can also react with the NO2 and NO in the system. Due to the presence of NOx in the system the formation of organic nitrates in small yields was also observed. It is interesting to compare the alkenes formed in the reactions of O(3P) with 1iodobutane and 2-iodobutane. In the reaction of O atoms with 1-iodobutane only one alkene can be formed, namely 1-butene, and this is observed with a yield of around 70%. In the case of 2-iodobtane, however, there is the possibility of the formation of both 1-butene and cis/trans -2-butene as the major alkene product (see Figure 11). The measurements show that the major alkene is 1-butene with a yield of around 46%. This results implies that adduct formation of the O atom between the I atom and the terminal methyl group is favoured over adduct formation between the O atom and the methylene group. Formation of trans-2-butene was observed, however, the yield was difficult to quantify and only a limit of <10% can be set at present. O H O + H3C

H

I

C

C

C

H

H

H

H H

H3C

H

I

C

C

C

H

H

H

CH3-CH2-CH=CH2

H

1-butene

+ HOI

1-iodobutane

hypoiodous acid

O H

I

H

C

C

C

H

H

H

I

H

H

C

C

C

H

H

H

major

H3C O + H3C

H

I

H

C

C

C

H

H

H

+ HOI

H O

2-iodobutane

H3C

Figure 11.

CH3-CH2-CH=CH2

H

H

minor

CH3-CH=CH-CH3 2-butene + HOI

Possible reaction pathways in the reaction of O(3P) with 1-iodobutane and 2iodobutane.

Conclusions The work has shown that a very small reactor can be used with a GC system to obtain kinetic data. Using the relative kinetic technique rate coefficients have been obtained for the reaction of O(3P) atoms with a series of alkyl iodides. Many of the rate coefficients are first time measurements. It has also been shown that alkenes and HOI are the major products of the reactions and the alkene has been quantified for the majority of the alkyl iodides studied. Acknowledgements This work was supported by the Bundesministerium für Bildung und Forschung within the “Atmosphärenforschung” programme (AFO 2000).

Kinetics, Products and Mechanism of O( 3P) Atom Reactions with Alkyl Iodides

205

References Barnes, I., Becker, K. H. and N. Mihalopoulos; An FTIR product study of the photooxidation of dimethyl disulfide, J. Atmos Chem., 18 (1994) 267-289. Barnes, I., K. H. Becker and T. Zhu; Near UV absorption spectra and photolysis products of difunctional organic nitrates, J. Atmos. Chem. 17 (1993) 353-373. Barnes, I., K. H. Becker, and S. Starke; FTIR spectroscopic observation of gaseous HOI, Chem. Phys. Lett. 196 (1992) 578-582. Carl, S. A. and J. N. Crowley; 298 K rate coefficients for the reaction of OH with i-C3H7I, n-C3H7I and C3H8, Atmos. Chem. Phys. 1 (2001) 1-7. Cavalli, F., H. Geiger, I. Barnes and K. H. Becker; FTIR kinetic, product, and modeling study of the OHinitiated oxidation of 1-butanol in air, Environ. Sci. Technol. 36 (2002) 1263-1270 Cotter, E. S. N., N. J. Booth, C. E. Canosa-Mas, D. J. Gray, D. E. Shallcross and R. P. Wayne; Reactions of Cl atoms with CH3I, C2H5I, 1-C3H7I, 2-C3H7I and CF3I: kinetics and atmospheric relevance, Phys. Chem. Chem. Phys., 3 (2001) 402–408. Cotter, S. A., C. E. Canosa-Mas, C. R. Manners, R. P. Wayne and D. E. Shallcross; Kinetic study of the reactions of OH with the simple alkyl iodides: CH3I, C2H5I, 1-C3H7I and 2-C3H7I, Atmos. Environ. 37 (2003) 11251133. Dillon, T. J. and D. E. Heard; A laser-induced fluorescence study of the O(3P) + CF3I and IO + NO reactions: Evidence for an association channel for O + CF3I at low temperatures, J. Photochem. Photobiol., 157 (2003) 223–230. Gilles, M. K., A.A. Turnipseed, R. K. Talukdar, Y. Rudich, P. W. Villalta, L. Gregory Huey, J. B. Burkholder and A. R. Ravishankara; Reactions of O(3P) with alkyl iodides: Rate coefficients and reaction products, J. Phys. Chem. 100 (1996) 14005-14015. Klaassen, J. J., J. Lindner and S. R.Leone; Observation of the Q1 OH(OD) stretch of HOI and DOI by Fourier transform infrared emission spectroscopy, J. Chem. Phys. 104 (1996) 7403-7411. Knyazev, V. D., V. S. Arutyunov and V. I. Vedeneev; The mechanism of O(3P) atom reactions with ethylene and other simple olefins, Int. J. Chem. Kinetics 24 (1992) 545-562. Loomis, R. A., J. J. Klaassen, J. Lindner, P. G. Christopher and S. R. Leone; Fourier transform infrared emission study of the mechanism and dynamics of HOI formed in the reaction of alkyl iodides with O(3P), J. Chem. Phys. 106 (1997) 3934-3947. Luo, D., J. A. Pierce, I. L. Malkina and W. P. L. Carter; Rate constants for the reactions of O(3P) with selected Monoterpenes, Int. J. Chem. Kinetics, 28 (1996) 1-8. Olariu, R. I., I. Barnes, K. H. Becker, and B. Klotz; Rate coefficients for the gas-phase reaction of OH radicals with selected dihydroxybenzenes and benzoquinones, Int. J. Chem. Kinet. 32 (2000) 696-702 Teruel, M. A., T. J. Dillon, A. Horowitz, and J. Crowley; Reaction of O(3P) with the alkyl iodides: CF3I, CH3I, CH2I2, C2H5I, 1-C3H7I and 2-C3H7I, Phys. Chem. Chem. Phys. 6 (2004) 2172-2178. Monks, P. S., L. J. Stief, D. C. Tardy, J. F. Liebman, Z. Zhang, S. –C. Kuo and R. B. Klemm; Discharge flowphotoionization mass spectrometric study of HOI: Photoionization efficiency spectrum and ionization energy, J. Phys Chem, 99 (1995) 16566-16570.

Kinetics of the Reaction between CF2 O and CH3OH Maximiliano A. Burgos Paci and Gustavo A. Argüello INFIQC, Departamento de Físico-Química, Facultad de Ciencias Químicas, Universidad Nacional de Córdoba, 5000 Córdoba, Argentina Key Words: Heterogeneous reactions, Fluorinated compounds, Kinetics, Methanol

Abstract The reaction between CH3OH and CF2O has been studied in the temperature range between 20 and 122 ºC. A considerable decrease in the reaction rate with increase in temperature was noted, indicating a contribution of heterogeneous character. Different types of surfaces and coatings have been used. The possible involvement of aerosols for the reaction to occur in the atmosphere is discussed. Introduction Methanol is a trace component of our atmosphere and significant concentrations (400-800 ppt) have been measured in the upper troposphere (Singh et al. 2000; Singh et al. 1995). A rough first inventory of the global sources of methanol indicates that its sources far exceed its known sinks (Singh et al. 2000). The large mismatch between the current atmospheric concentration and the estimated sources and sinks of this oxygenated species indicates that substantial and still unknown removal processes, other than removal with OH radicals, must exist. The existence of unknown, possibly heterogeneous processes, which could provide effective sinks for methanol, cannot be currently ruled out (Singh et al. 2000). Removal reactions for this species must, therefore, be found within the atmospheric constituents, and the search should not only be confined to homogeneous processes, but should also be extended to heterogeneous processes. We have been exploring many different atmospheric constituents as reaction partners for CH3OH in the gas phase using an in situ FTIR experimental setup to follow the course of the reactions. Among many very well-known substances, there is a species, whose concentration has been growing steadily in the atmosphere because it comes from the degradation of almost any fluorinated molecule released into the atmosphere, i.e. carbonyl fluoride (CF2O). CF2O could represent almost up to a third of the total inorganic fluorine in the stratosphere (Raper et al. 1987). We have found that CF2O reacts in the gas phase, though with a very small rate constant which makes the homogeneous process almost unimportant, in contrast the heterogeneous contribution of this reaction is very important. The relative importance of the heterogeneous contribution was studied using smallvolume reaction chambers, where wall effects were evident and allowed different surfaces and surface conditions to be investigated. We have previously demonstrated the active role played by the surface in the handling of fluorinated species (Pernice, 2002). CF2O, in particular, is considered one of the most stable end products in the atmospheric degradation of almost any CFC, HCFC, HFC or HFE because of its strong bonds that allow this molecule to survive most of the radiation found below 30 km of altitude without being degraded. Nevertheless, it reacts with

207 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 207–212. © 2006 Springer. Printed in the Netherlands.

208

M. A. Burgos Paci and G. A. Argüello

CH3OH, gives a unique product (the characterization of which is in progress) irrespective of the type of reaction (homogeneous or heterogeneous) and the rate of the reaction increases with decreasing temperature. This temperature behaviour is particularly interesting because the temperature profile of the atmosphere and the vertical distribution of CF2O favour (in the presence of particulate matter) the disappearance of methanol. Experimental Volatile compounds were manipulated in a conventional vacuum line with three U-traps, PTFS stop-cock valves (Young, London) and two pressure gauges (Bell and Howell). The line was connected directly to the reaction cell, which lies in the optical axis of an FTIR spectrometer (Bruker IFS28), in a configuration fairly similar to the one used in smog chamber experiments. This set-up allowed the timely recording of IR spectra and all the analyses were therefore performed by FTIR spectroscopy. The reactions have been carried out by mixing different pressures of CH3OH (1 – 30 mbar) and CF2O (1 – 25 mbar) in the reaction cell. A double-walled reaction cell connected to a thermostat was used to investigate the effect of temperature. Water and glycerine were used as heating liquids for temperatures between 20 – 90 and 90 – 122 ºC, respectively. In order to vary the type of surface the reactants were exposed to, three reaction cells were employed: a) a glass cell with and without small glass balls added, b) a stainless steel cell, and c) a glass cell with the inside walls covered with an inert wax (Halocarbon). In order to cover the glass cell with Halocarbon wax, it was first spread on the inside walls. The cell was then heated in an oven at 70 ºC for approximately 1 hour until the wax formed a film over the walls. Results and Discussion In order to follow the evolution of the system as the reaction takes place, IR spectra were recorded in all the experiments immediately after the reactants were mixed. Figure 1 shows IR spectra acquired before (A), after 15 min (B) and at the end (C) of the reaction of 18 mbar CH3OH and 3 mbar CF2O. The disappearance of CF2O was monitored using its PQR band at 1985 cm-1. The rate of disappearance was always faster than that of CH3OH. The time evolution of CH3OH was followed at 1095 cm-1 though it was quite difficult to quantify because of the interference of SiF4. The reaction results in the formation of a still unidentified product with characteristic IR bands at 1850 cm-1, which is typical of a C=O stretch, and other bands at 1250 cm-1 (C—F stretch), 1455 cm-1 (CH bending), 915 cm-1 and 785 cm-1 which are shown in panel (D) of Figure 1. Several experimental runs with different initial concentrations of each reactant were carried out to investigate the effect in the rate of product formation. The curves obtained were fitted in all cases with the equation [Prod] = Ao[1-exp(-k t)]. Plot A in Figure 2 shows the case where the initial pressure of CH3OH was fixed (8.0-9.0 mbar) and the CF2O initial pressure was set at 0.8; 10.0 and 11.0 mbar. It is worth noting that the amount of product formed increases as the initial pressure of CF2O is increased, i.e. it is much smaller for an initial pressure of 0.8 mbar than for one of 10.0 mbar. Another feature to notice is that when the initial pressure of CF2O is 10.0 – 11.0 mbar, and 0.8 mbar the stationary product concentration is reached at around 50 min. and 200 min. respectively. This indicates that the rate of the reaction increases as the initial

Kinetics of the Reaction between CF2 O and CH3 OH

209

pressure of CF2O increases. Analogous observations can be made when the initial pressure of CF2O is fixed and that of CH3OH is varied.

6

D

Abs(a.u.)

5 4

C

3

B

2 1

A

0 3000

2500

2000

cm Figure 1.

-1

1500

1000

500

Sequence of IR spectra obtained at different times of reaction.

The order of the rate with respect to CF2O was obtained by running the reaction with an excess of CH3OH and conversely, to obtain the order with respect to the concentration of CH3OH at least a ten fold excess of CF2O was used. Under these conditions the logarithm of the concentration of the limiting reagent as a function of time can be excellently fitted with a linear function, as shown in Plot B of Figure 2 for the case of CF2O. The order of the reaction is one with respect to each reagent and order two overall. Under the experimental conditions the pseudo first order constant for the disappearance of CF2O at 25 ºC was 2.5x10-3 s-1. Effect of temperature The temperature effect on the rate of reaction was studied between 20 and 122 ºC. Plot A of Figure 3 shows the formation of the main still unidentified product in three experiments performed at 25, 100 and 122 ºC with the same initial pressures of reactants. A decrease in the product formation rate was observed as the temperature increased. It is interesting to notice that at all temperatures the products are the same. The decrease of the rate at higher temperatures can be rationalized in terms of a heterogeneous reaction taking place on the surface of the cell, since as the temperature goes up there should be a correspondingly higher percentage of free gaseous molecules not adsorbed onto the walls that should react homogeneously.

M. A. Burgos Paci and G. A. Argüello

210

Presently, we are trying to determine whether the reaction could occur on different surfaces and whether it could give other products. 0.4

-0.5

A

0.3

B

-1.0

0.2

9mb CH3OH + 11.0mb CF2O 8mb CH3OH + 10.0mb CF2O 8mb CH3OH + 0.8mb CF2O

0.1

ln([CF2O])

Abs (a.u.)

-1.5 -2.0 -2.5 -3.0 -3.5 0.0

-4.0 0

50

100

150

time (min.)

Figure 2.

200

250

0

5

10

15

20

time (min)

A) Effect of the pressure of CF2O on product formation. B) Pseudo first order linear fit for the decomposition of CF2O.

Effect of different surfaces In order to investigate the effect of the surface with which the reactants collide, experiments were performed to change the surface of the cell. Plot B of Figure 3 shows the effect of using a glass (squares in the figure) or a stainless steel (circles) reaction vessel. As can be seen the type of surface clearly affects the reaction rate. Though the experimental runs did not reproduce exactly the same pressures, the differences in rate are beyond any reasonable doubt. It is observed that in the glass cell, the rate is faster than in the stainless steel cell. This observation can be taken as a proof of the catalytic effect of glass on the reaction. The ratio of the constants for the formation of product (taken from the equations [Prod] = Ao[1-exp(-k t)]) indicates that the rate is 20 times higher in the glass cell. A hypothesis could be that the OH radicals and silica groups on the glass surface facilitate the adsorption of CH3OH by dipole-dipole interactions thus providing the proper orientation of the molecule to be attacked or the weakening of the bond that will constitute the reaction coordinate. In the stainless steel cell there is no adsorption of CH3OH; thus the catalytic effect is reduced and the reaction should proceed almost entirely in the gas phase giving the same product as in the glass cell. Plot C in Figure 3 shows three experiments: one using a glass cell, another with a glass cell with glass beads added to increase the surface, and a third one employing a cell whose walls were covered with wax. From a comparison of the first two experiments (squares and triangles) it can be seen that the addition of glass surface does not have a quantitative effect on the reaction rate. This could be because the amount of glass in the walls of the cell far exceeds the active surface needed to catalyze the reaction and so the added glass balls do not increase the reaction rate. The diamonds in the figure correspond to the waxed glass cell. Here, an induction time is observed after which, the reaction proceeds as in the “normal” glass cell in terms of the amount of product formed. We believe that in the cell there may have been some points where the wax film was too thin and after the first 30 min of the reaction some molecules could have reached the

Kinetics of the Reaction between CF2 O and CH3 OH

211

glass surface; once this occurred the catalytic effect of glass accelerated the reaction. This reinforces the idea that a small amount of active surface is needed to catalyze the reaction. 0.40 0.35 0.30

Abs (a.u.)

0.25 0.20 0.15 0.10

A

0.05 0.00 0

10

20

30

40

50

60

time (min.)

0.40 0.5

0.35

0.4

0.25

Abs (a.u.)

Abs (a.u.)

0.30

0.20 0.15 0.10

0.2 0.1

0.05

B

0.00 0

60

120

180

240

time (min.)

Figure 3.

0.3

300

360

420

C

0.0 0

30

60

90

120

150

180

210

time (min.)

A) Temperature effect on the rate of product formation. Ŷ T=25 ºC; Ɣ T=100 ºC; Ÿ T=122 ºC. B) Product formation when the reaction is carried out in a glass (Ŷ) and a stainless steel (Ɣ) cell. C) Product formation when the reaction is carried out in a glass cell (Ŷ), a glass cell with glass beads (Ÿ) and a waxed cell (Ƈ).

Atmospheric Implications This preliminary study of the reaction between CF2O and CH3OH shows that it has an important heterogeneous component, giving a mean lifetime of 6 min. at 25 ºC for CH3OH when the reaction is carried out with an active surface. We think that the homogeneous gas phase reaction is too slow to be considered as an important process since it is 20 times slower than the heterogeneous reaction. Thus taking into account that: a) the reaction is catalyzed heterogeneously and in the atmosphere there is a high possibility of having heterogeneous processes due to the presence of aerosols, particulate matter, water droplets, etc., and b) the reaction rate increases as the temperature decreases and in the upper troposphere temperatures

212

M. A. Burgos Paci and G. A. Argüello

decrease to a mean value of -50 ºC. This reaction could have an important role in explaining the low concentration of CH3OH in the atmosphere. However, further work is needed to assess the feasibility of this hypothesis. Problems to solve are: i) the nature of the product of the reaction and its associated chemistry and photochemistry, ii) the stoichiometry of the reaction, and iii) the effect of atmospheric aerosols on the reaction rate. Acknowledgements The authors thank CONICET, SECyT and ANPCyT for financial support. M.A.B.P. thanks CONICET for a postdoctoral fellowship. Language assistance by Karina Plasencia is gratefully acknowledged. References Singh H., Y. Chen, A. Tabazadeh, Y. Fukui, I. Bey, R. Yantosca, D. Jacob, F. Arnold, K. Wohlfrom, E. Atlas, F. Flocke, D. R. Blake, N. Blake, B. Heikes, J. Snow, R. Talbot, G. Gregory, G. Sachse, S. Vay and Y. J. Kondo; Distribution and fate of selected oxygenated organic species in the troposphere and lower stratosphere over the Atlanatic, Geophys. Res. 105 (2000) 3795-3805. Singh H., M. Kanakidou P. J. Crutzen, and D. J. Jacob; High concentrations and photochemical fate of oxygenated hydrocarbons in the global troposphere, Nature 378 (1995) 50-54. Pernice H., H. Willner, K. L. Bierbrauer, M.Burgos Paci and G. A. Argüello; Fluoroformic Acid Anhydride, FC(O)OC(O)F, Angew. Chem. Int. Ed. 41 (20) (2002) 3832-3834. Raper O. F., C. B. Farmer, R. Zander and J. H. Park, Ground-Based Infrared Measurements of Tropospheric Source Gases Over Antarctica During the 1986 Austral Spring, J. Geophys. Res. 92 (1987) 9851-9856.

New Kinetic and Spectroscopic Measurements in the CF3Ox + NOx System M. S. Chiappero1 , F. E. Malanca1 , G. A. Argüello1 S. Nishida , K. Takahashi2, Y. Matsumi2, M. D. Hurley3, and T. J. Wallington3 2

1

INFIQC, Departamento de Físico-Química, Facultad de Ciencias Químicas, Universidad Nacional de Córdoba, 5000 Córdoba, Argentina 2 Solar-Terrestrial Environment Laboratory and Graduate School of Science, Nagoya University, Toyokawa, Aichi 442-8507, Japan 3 Department of Physical and Environment Sciences, Ford Research and Advanced Engineering, SRL-3083, Ford Motor Company, Dearborn, MI 48121-2053, USA Key Words: Fluorinated peroxy radicals, Kinetic, Nitric Oxides, Organic nitrates, UV Spectroscopy

Abstract CF3Ox radicals are present in the atmosphere as oxidized species of a series of CFC´s, HFC´s and HCFC´s. They react with NOx species. In particular, the reaction between CF3O2 and NO which shows two reaction channels, CF3O2 + NO o CF3O + NO2

k1a

(1a)

CF3O2 + NO + M o CF3ONO2 + M

k1b

(1b)

has been revised using FTIR smog chamber techniques to provide the first systematic study of the nitrate yield thus allowing the determination of the branching ratio for both processes. In 700 Torr of N2/O2 diluent at 296 K the branching ratio is k1b/(k1a+k1b) = (1.67±0.27)×10-2 and the CF3ONO2 forming channel appears to approach the high-pressure limit for pressures above 100 Torr. This finding could probably provide more conventional energy transfer parameters to account for the pressure dependence in atmospheric models. CF3ONO2 is also the main product observed in the black lamp (Omax|355 nm) photolysis of pure CF3O2NO2. The thermal decomposition of the peroxide (which does not absorb significantly beyond 300 nm) leads to an equilibrium with CF3O2 and NO2 which does absorb and thus forms NO. Subsequent reactions lead to CF3O radicals, that in presence of NO2 form CF3ONO2, giving the overall reaction: CF3O2NO2 + hQ o CF3ONO2 + ½ O2

(2)

The low branching ratio in the formation of CF3ONO2, the photochemical characteristics of CF3O2NO2 just described and its probable even distribution in the free troposphere, led us to look closer at the UV features of the peroxide. The UV spectrum at room temperature was already known. We have determined the temperature dependence between 225 and 300 K. The results show an increase in the cross sections with temperature. From the data, photochemical half-lives in the atmosphere from 20 to 450 days result, depending on altitude and solar zenith angle. In terms of the process controlling CF3O2NO2 decomposition, calculations show that from the surface to around 2 km altitude, the thermal process is faster than photolysis. Beyond 2 km, photochemistry governs the decomposition for low and mid latitudes. At higher latitudes (50 degrees) the limiting process below the tropopause is thermal decay. 213 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 213–221. © 2006 Springer. Printed in the Netherlands.

M. S. Chiappero et al.

214 Introduction

In the atmosphere, the CFC´s, HFC´s and HCFC´s containing the CF3 moiety produce CF3Ox radicals that react with NOx. The reaction between CF3O2 and NO shows two possible channels: CF3O2 + NO o CF3O + NO2

(1a)

CF3O2 + NO + M o CF3ONO2 + M

(1b)

Channel (1a) converts NO into NO2 thereby leading to net O3 formation since the production of tropospheric ozone is due mainly to NO2 photolysis. NO2 + hQ (O < 420 nm) o NO + O(3P)

(3)

O(3P) + O2 + M o O3 + M

(4)

Reaction channel (1b) produces alkyl nitrates which are less reactive than either NO or NO2 and hence limits the photochemical formation of ozone. The formation of organic nitrate via reaction (1b) provides a tracer of photochemical ozone production in the atmosphere (Bertman et al., 1995; Flocke et al., 1991; Flocke et al., 1994; Roberts et al., 1998). Therefore, an understanding of the processes which control tropospheric ozone levels is essential in assessments of the global oxidizing capacity of the atmosphere and the anthropogenic contribution to global climate change. There are no literature data available concerning the branching ratio k1b/(k1a+k1b) and therefore no estimates of the importance of nitrate formation in reaction (1). Furthermore, there have been no direct studies of the nitrate yield for any halogenated peroxy radical. The organic nitrate is an inactive reservoir species for NOx. These reservoirs may be subject to dry and wet deposition (Treves et al., 2000), as well as thermal and/or photochemical decomposition. Peroxynitrates, RO2NO2, are thermally unstable and decompose to reform RO2 radicals and NO2. At typical temperatures of the lower troposphere the half life of peroxynitrates with respect to thermal decomposition is of the order of minutes whilst at the colder temperatures typical of the upper troposphere these compounds have half lives measured in months or years. Acting as reservoir compounds for NOx, they can transport NOx (and RO2 radicals) over long distances. The fluorinated organic nitrates CF3ONO2 and CF3O2NO2 have been well characterized (Sander et al., 2001; Kopitzky et al., 1998; Hohorst and DesMarteau, 1974) and show UV absorption below 300 nm. CF3O2NO2 absorbs significantly in the wavelength region 200-300 nm (Kopitzky et al., 1998). Photolysis is thus a potential loss mechanism for CF3O2NO2 in the stratosphere which needs to be considered. UV absorption cross section measurements as a function of temperature are also needed as inputs to global atmospheric chemistry programs in order to calculate photolysis frequencies. The majority of such currently available measurements are related to primary pollutants such as the CFCs (Hubrich et al., 1977; Simon et al., 1988; Merienne et al., 1990); HCFCs (Robbins, 1976; Vanlaethem-Meuree et al., 1978); HFEs (Orkin et al., 1999); and alkyl iodides (Roehl et al., 1997; Rattigan et al., 1997; Fahr et al., 1995). There have been comparatively few studies of the UV spectra of stable primary oxidation products of halogenated compounds (e.g., CF3C(O)Cl, CF3C(O)F, and CF3C(O)H, (Rattigan et al., 1993)).

New Kinetic and Spectroscopic Measurements in the CF3Ox + NOx System

215

The data base for thermally unstable products such as CF3C(O)O2NO2 (perfluoroperoxyacetylnitrate) (Libuda et al., 1995), and CF3O2NO2 is particularly sparse. For the reasons referred to above, the aims of this work are (i) the study of the branching ratio k1b/(k1a+k1b) to elucidate the factors influencing the CF3ONO2 yield in the reaction of CF3O2 radicals with NO, (ii) the study of the photochemical (O>300 nm) behaviour of CF3O2NO2 and (iii) the study of the influence of temperature on the UV absorption features of CF3O2NO2. Experimental

The study of the reaction between CF3O2 and NO at 296 ± 2 K, was carried out with different mixtures of CF3N2CF3 (6.18–11.8 mTorr), NO (14.7–241 mTorr) and O2 (19.2–100 Torr ) in N2 diluent (20–700 Torr) that were admitted into a 140 L Pyrex reactor interfaced to a Mattson Sirus 100 FTIR spectrometer. The reactor was surrounded by 22 fluorescent blacklamps (Wallington and Japar, 1989). Each mixture was subjected to UV irradiation and the time evolution of the system was followed by FTIR spectroscopy (Wallington and Ball, 1995a). CF3O2 radicals were produced in situ by the reaction of O2 with the CF3 radicals formed by photolyis of CF3N2CF3 (Omax<355 nm). Control experiments were performed in which CF3N2CF3/NO/O2/N2 mixtures as well as their products were allowed to stand in the dark in the chamber for 30 min. In other experiments, the effect of black lamp radiation (GE FL6 BLB) on CF3O2NO2 was studied using a 23 cm long quartz cell fitted at both ends with KBr windows. The quartz cell was mounted in the sample compartment of a Bruker IFS28 FTIR spectrometer (Bruker, Karlsruhe, Germany) equipped with a DTGS detector and KBr beam splitter. All experiments were performed at 298 K. Spectra were obtained by co-adding 8 interferograms in the range 4000 to 400 cm-1 with 2 cm-1 resolution. In addition, the dependence of the UV absorption with temperature was studied using a diode array UV spectrometer with a standard 10 cm quartz gas cell mounted and aligned inside an evacuated metal box with two opposed quartz windows. The body of the cell was surrounded with a plastic coil connected to an external thermostat-cryostat for temperature control (300-220 K). The temperature of the gas within the cell could be regulated to within 0.1oC. For atmospheric half-life calculations, the program GCSOLAR release 1.2 (Zeep et al.,1997) with GCSOLAR ELEVATION (which allows to compute photolysis rate constants of a particular substance as function of elevation above sea level), was used. Unity quantum yield, a terrestrial atmosphere and Greenwich meridian, were assumed for all latitudes. Reagents The reagents CF3N2CF3, CF3O2NO2, and CF3ONO2 were obtained as described below: CF3N2CF3 was synthesized by passing cyanogen chloride diluted with nitrogen or helium, over a bed of silver fluoride (AgF2) and collecting the product in a cold trap (pentane/liquid nitrogen or liquid nitrogen) as described elsewhere (Wallington and Ball, 1995b).

216

M. S. Chiappero et al.

CF3O2NO2 was synthesized by mixing (CF3CO)2O, NO, and O2 in a 12 L reactor, maintained at -25 ºC and subjecting the mixture to 254 nm radiation (Burgos Paci et al., 2003). CF3O2NO2 was purified by fractional distillation under vacuum at -100/-120/-196 ºC. The peroxide was obtained in the -120 ºC fraction (Chiappero et al., 2004). CF3ONO2 was synthesized photolysing a pure sample of CF3O2NO2 with a black lamp (Omax|355 nm). The product was purified by trap to trap distillation (Chiappero et al, 2004). Other reagents (N2, O2, and NO) were obtained from commercial sources at purities >99% and were used as received. Results and Discussion The photolysis of CF3N2CF3 in the presence of O2 leads, through the sequence: CF3N2CF3 + hQ o 2 CF3 + N2

(5)

CF3 + O2 + M o CF3O2 + M

(6)

to the formation of CF3O2 radicals. In the presence of excess NO (14.7-241 mTorr) the fate of CF3O2 radicals is dominated by reaction (1). COF2 and CF3ONO2 were the only observed carbon-containing products. Within the experimental uncertainties the observed formation of COF2 and CF3ONO2 accounts for 100% of the loss of the CF3N2CF3. The observed formation of CF3ONO2 provides information concerning k1b/(k1a+k1b). Assuming that reaction (1) is the sole loss of CF3O radicals then the branching ratio is k1a/k1b ='[COF2]/'[CF3ONO2].

' [COF2] (mTorr)

10

5 Total Total Total Total Total

0 0.0

pressure pressure pressure pressure pressure

0.1

= = = = =

20 Torr 50 Torr 100 Torr 300 Torr 700 Torr

0.2

' [CF 3 ONO 2 ] (mTorr)

Figure 1.

Formation of COF2 versus CF3ONO2 for experiments conducted at selected pressures. The lines plotted correspond to linear least square fits. The solid line comprises the data points for 100, 300, and 700 Torr and has a slope = 0.0167 (see text for details).

New Kinetic and Spectroscopic Measurements in the CF3Ox + NOx System

217

As the branching ratio k1b/(k1a+k1b) is expected to display a dependence on total pressure due to the termolecular nature of reaction (1b), experiments were carried out at different total pressures (N2/O2 as diluent). As seen from Figure 1, within the admittedly significant data scatter there is no discernable effect of total pressure over the range 100–700 Torr on the relative yields of COF2 and CF3ONO2. However, the [COF2]/[CF3ONO2] ratio observed at 20 and 50 Torr total pressure is significantly different from that at 100–700 Torr. The ratio k1b/(k1a+k1b) has values of 3.5 ×10-3, 7.6 ×10-3 and 1.67 ×10-2 for total N2/O2 diluent pressures of 20, 50 and 100-700 Torr and 296 K, respectively. The simplest explanation of these observations is that the rate of reaction (1b) is close to its high-pressure limiting value for pressures above 100 Torr but in the fall off regime at pressures of 50 Torr and below. It is likely that the pressure dependence reported herein for reaction (1b) can be modelled using more conventional energy transfer parameters as Barker and Golden (2003) suggest. The effect of the UV radiation (Omax|355 nm) over a room temperature equilibrated CF3O2NO2/NO2 mixture (CF3O2NO2 2.6 mbar) can be seen in Figure 2. This figure shows IR spectra before and after UV irradiation. IR features attributable to CF3O2NO2 at 792, 1192, 1244, 1303, and 1762 cm-1 (Kopitzky et al., 1998) are visible in panel A.

2 .0 1 .5 1 .0

A : b e fo re U V

0 .5 0 .0

Absorbance

1 .5 1 .0

B : 1 5 m in u te s U V

0 .5 0 .0

1 .5 1 .0

C : 3 0 m in u te s U V

0 .5 0 .0

1 .5 1 .0

D : 4 0 m in u te s U V

0 .5 0 .0

600

800

1000

1200

1400

1600

1800

2000

W a v e n u m b e r (c m -1 )

Figure 2.

IR spectra acquired before (A) and after (B), (C) and (D) UV irradiation. It can be seen the decrease in CF3O2NO2 concentration and the appearance of peaks at 788, 1156, 1243, 1265 and 1748 cm-1 which can be assigned to CF3 ONO2 (Sander et al., 2001). The bands at 1944 cm-1 correspond to COF2.

The results could be interpreted as follows: in the mixture CF3O2NO2/NO2, the peroxide (which does not absorb significantly beyond 300 nm) (Kopitzky et al., 1998) exists in equilibrium with CF3O2 radicals and NO2 CF3O2NO2 + M o CF3O2 + NO2 + M

(7)

CF3O2 + NO2 + M o CF3O2NO2 + M

(-7)

218

M. S. Chiappero et al.

and the irradiation of the mixture converts NO2 into NO (reaction 3) (Crutzen, 1973), which is then oxidized back to NO2 via reaction with CF3O2 radicals (reaction 1a). Subsequently, the resulting CF3O radicals react with NO2 to give CF3ONO2 CF3O + NO2 + M o CF3ONO2 + M

(8)

The result is photochemical conversion of CF3O2NO2 into CF3ONO2 in high yield with an overall stoichiometry of CF3O2NO2 o CF3ONO2 + ½ O2

(2)

The perfluoromethyl peroxinitrate has a dissociative UV spectrum below 300 nm similar to those of other peroxynitrates. However, as noted previously (Kopitzky et al., 1998), it is approximately a factor of 2 less intense (Morel et al., 1980). Table 1 shows the temperature dependence of the UV cross sections of CF3O2NO2 at 233, 253, 273, 294 and 300 K. As the peroxynitrate exists in equilibrium with NO2, and this species in turn exists in equilibrium with N2O4, at lower temperatures a greater fraction of the NO2 exists in the dimer form N2O4. NO2 and N2O4 absorb in the wavelength range studied (N2O4 being a particularly strong absorber) and the features of these two substances must be subtracted from the experimental spectra. This stripping process was repeated with the measurements at low temperature taking into account the variation in concentrations due to the change in the value of the equilibrium constant. The increase in absorption cross section with temperature seen from examination of Table 1 is consistent with an increased population of the vibrational and rotational levels of the ground electronic state of the molecule at higher temperatures (Okabe, 1978). In the troposphere, the temperature decreases with altitude increase up to the tropopause (approximately 10 km altitude). In this zone the temperature is close to 220 K and precisely, under these conditions, the peroxynitrate UV absorption cross sections are the least. This means that its photodissociation rate will be lower than at the surface. The precise value integrated for the whole absorption region will depend, of course, on the solar flux, that is richer in short wavelength radiation than the surface flux. In fact, a detailed calculation of the CF3O2NO2 half life and its dependence with altitude and latitude shows a maximum half life at around 15 km. This maximum reaches values as long as 220 days for mid-latitudes allowing the compound to be transported though the atmosphere. Therefore, CF3O2NO2 could be transported either through convective stream to the stratosphere, or to remote tropospheric places where it can dissociate again. From the surface up to 2 km, the thermal decomposition loss (Mayer-Figge et al., 1996) is always faster than the photochemical loss. Beyond this height the trend is opposite at low and mid latitudes and photochemistry governs the loss. At high latitudes, the photochemical half life is again longer. The use of Kopitzky´s cross sections, with no temperature dependence, results in a steady decrease of the half life with altitude. Furthermore, the absolute values are all shorter than 20 days; therefore our temperature dependent cross sections could help to better understand the fate of CF3O2NO2 since the half life is in any case considerablely longer than thought before.

New Kinetic and Spectroscopic Measurements in the CF3Ox + NOx System Table 1.

219

CF3O2NO2 cross-sections (in units of 10-20 cm2 molecule-1) at different temperatures. Values are averaged over 5 nm intervals. Values printed in italics are derived from absorbance readings close to the baseline drift, therefore, their uncertainties are high. CF3O2NO2 cross-sections

 O (nm) 200 205 210 215 220 225 230 235 240 245 250 255 260 265 270 275 280 285 290 295 300 305 310 315 320 325 330 335 340

Kopitzky 301 K 220 150 120 74 54 46 38 32 27 22 18 14 11 8.2 5.8 4.0 2.6 1.7 1.0 0.69 0.45 0.28 0.20 0.13 0.08 0.05 0.037 0.024 0.014

300 K 215 149 115 77 58 47 40 35 29.5 24.0 19.0 14.7 11.5 8.5 5.8 3.96 2.76 1.85 1.31 0.88 0.60 0.45 0.31 0.23 0.16 0.12 0.096 0.040 0.029

294 K 215 149 111 76 58 47 40 35 29.3 23.8 18.9 14.8 11.2 8.2 5.8 3.94 2.65 1.75 1.18 0.76 0.5 0.35 0.23 0.16 0.11 0.071 0.048 0.033 0.022

This Work 273 K 216 150 114 77 57 47 39 33 27.7 22.3 17.6 13.6 10.1 7.3 5.1 3.32 2.15 1.35 0.86 0.54 0.32 0.22 0.14 0.086 0.054 0.034 0.022 0.014 0.009

253 K 216 150 115 78 56 46 38 32 26.6 21.3 16.8 12.8 9.5 6.8 4.6 2.96 1.84 1.11 0.66 0.4 0.22 0.14 0.085 0.051 0.03 0.018 0.011 0.007 0.004

233 K 215 149 114 77 55 45 37 30 25.7 20.5 16.2 12.3 9.1 6.4 4.3 2.68 1.57 0.87 0.45 0.23 0.11 0.058 0.03 0.015 0.008 0.003 0.001 -

Conclusions FTIR smog chamber techniques have been used successfully to show that reaction of CF3O2 radicals with NO proceeds via two channels giving either CF3O + NO2 (1a) or CF3ONO2 (1b) as products. The branching ratio for CF3ONO2 formation has been determined to be, k1b/(k1a+k1b) = (1.67 ± 0.27) × 10

-2

in 700 Torr of N2/O2 diluent at 296 K. Reaction (1b) appears to be close to its high pressure limiting value for pressures above 100 Torr.

220

M. S. Chiappero et al.

CF3O2NO2 reacts both by photochemical and thermal processes. The photolysis of a mixture in equilibrium of CF3O2NO2 at Omax|355 nm radiation leads to the formation of CF3ONO2 due mainly to the photolysis of NO2. There is a clear dependence of the cross sections with temperature. Although the photochemical half-life increases with latitude, the decomposition is mainly thermal within the boundary layer. If the temperature dependence in the cross sections were not taken into account, calculations would predict a half-life governed solely by photolysis. Acknowledgements T.J.W. and M.S.C. thank the National Science Foundation for a WISC grant. The Nagoya group thanks both Grants-in-Aid from the Ministry of Education, Science, and Culture Financial and also the Sumito Foundation for financial support. M.S.C., F.E.M. and GAA thank to CONICET, SeCyT-UNC, Antorchas for financial support. References Barker J.R. and D.M. Golden; Master Equation Analysis of Pressure-Dependent Atmospheric Reactions, Chem. Rev. 103 (2003) 4577-4592. Bertman S.B., J.M. Roberts and H. Westberg; Evolution of alkyl nitrates with air mass age, J. Geophys. Res., 100 (1995) 22805-22814. Chiappero M.S., M.A. Burgos Paci, G.A. Argüello and Wallington T.J.; Novel synthesis of trifluoromethylnitrate, CF3ONO2, Inorg. Chem. 43, (2004) 2714-2716. Crutzen P.J.; A discussion of the chemistry of some minor constituents in the stratosphere and troposphere, Pure Appl. Geophys. 106 (1973) 1385-1399. Fahr A., A.K. Nayak and R.E. Huie; Temperature dependence of the ultraviolet absorption cross section of CF3I, Chem. Phys. 199 (1995) 275-284. Flocke F., A. Volz-Thomas and D. Kley; Measurements of alkyl nitrates in rural and polluted air masses, Atmos. Environ. 25A (1991) 1951-1960. Flocke, F., A. Volz-Thomas and D. Kley; The use of alkyl nitrate measurements for the characterization of the ozone balance at TOR-Station N. 11, Schauinsland, in: P. B. Borrell, P. M. Borrell, T. Cvitas, and W. Seiler (eds), Transport and transformation of pollutants in the troposphere - EUROTRAC Symposium, SPB Academic Publishing (1994) 243-247. Hohorst F.A. and D.D. DesMarteau; Some reactions of CF3OO derivatives with inorganic nitrogen compounds. Synthesis and vibrational spectrum of trifluoromethyl peroxynitrate, Inorg. Chem. 13 (1974) 715-719. Hubrich C., C. Zetsch and F. Stuhl; Absorptionsspektren von halogenierten methanen im bereich von 275 bis 160 nm bei temperaturen von 298 und 208 K, Ber. Bunsenges. 81 (1977) 437- 442. Kopitzky R., H. Willner, H.-G. Mack, A. Pfeiffer and H. Oberhammer; IR and UV absorption cross sections, vibrational analysis, and the molecular structure of trifluoromethyl peroxynitrate, CF3OONO2, Inorg. Chem. 37 (1998) 6208-6213. Libuda H.G. and F. Zabel; UV absorption cross sections of acetyl peroxynitrate and trifluoroacetyl peroxynitrate at 298 K, Ber. Bunsenges. Phys. Chem. 99 (1995) 1205-1213. Merienne M.F., B. Coquart and A. Jenouvrier; Temperature effect on the ultraviolet absorption of CFCl3, CF2Cl2 and N2O, Planet. Space Sci. 38 (1990) 617-625. Mayer-Figge A., F. Zabel and K.H. Becker; Thermal decomposition of CF3O2NO2, J. Phys. Chem. 100 (1996) 6587-6593. Morel O., R. Simonaitis and J. Heicklen; Ultraviolet absorption spectra of HO2NO2, CCl3O2NO2, CCl2FO2NO2, and CH3O2NO2, Chem. Phys. Lett. 73 (1980) 38-42.

New Kinetic and Spectroscopic Measurements in the CF3Ox + NOx System

221

Okabe H.; Photochemistry of small molecules; (1978) p.p. 18-20. John Wiley & Sons, New York. Orkin V.L., E. Villenave, R.E. Huie and M.J. Kurylo; Atmospheric lifetimes and global warming potentials of hydrofluoroethers: Reactivity toward OH, UV spectra, and IR absorption cross sections, J. Phys. Chem., A. 103 (1999) 9770-9779. Rattigan O.V., O. Wild, R.L. Jones and R.A. Cox; Temperature-dependent absorption cross-sections of CF3COCl, CF3COF, CH3COF, CCl3CHO and CF3COOH, J. Photochem. Photobiol. A: Chem. 73 (1993) 1-9. Rattigan O.V., D.E. Shallcross and R.A. Cox; UV absorption cross-sections and atmospheric photolysis rate of CF3I, CH3I, C2H5I and CH2Icl, J. Chem. Soc. Faraday Trans. 93 (1997) 2839-2846. Roberts J. M., S.B. Bertman, D.D. Parish, F.C. Fehsenfeld, B.T. Jobson and H. Niki; Measurement of alkyl nitrates at Chebogue Point, Nova Scotia during the 1993 North Atlantic Regional Experiment (NARE) intensive, J. Geophys. Res., 103 (1998) 13569-13580. Robbins D.E.; Photodissociation of methyl chloride and methyl bromide in the atmosphere, Geophys. Res. Lett. 3 (1976) 213-216. Sander S., H. Willner, H. Oberhammer and G.A. Argüello; Trifluormethylnitrat, CF3ONO2, Z. Anorg. Allg. Chem. 627 (2001) 655-661. Simon P.C., D. Gillotay, N. Vanlaethem-Meuree and J. Wisemberg; Ultraviolet absorption cross-sections of chloro and chlorofluoro-methanes at stratospheric temperatures, Atmos. Chem. 7 (1988) 107-135. Treves K., L. Shragina and Y. Rudich; Henry’s law constants of some E-, J-, and G-hydroxyalkyl nitrates of atmospheric interest, Environ. Sci. Technol. 34 (2000)1197-1203. Vanlaethem-Meuree N., J. Wisemberg and P.C. Simon; Influence de la temperature sur les sections efficases d'absorption des chlorofluoromethanes dans l'ultraviolet, Bull. Acad. Roy. Belgique, Cl. Sci. 64 (1978) 31. Wallington T.J. and S.M. Japar; Fourier Transform Infrared Kinetic Studies of the Reaction of HONO with HNO3, NO3, and N2O5 at 295K, J. Atmos. Chem., 9 (1989) 399-409. Wallington, T.J. and Ball, J.C.; Atmospheric chemistry of CF3O radicals: reaction with O3. Chem. Phys. Lett. 234 (1995a) 187-194. Wallington T. J. and J.C. Ball; Atmospheric chemistry of CF3O radicals: reaction with CH4, CD4, CH3F, CF3H, 13 CO, C2H5F, C2D6, C2H6, CH3OH, i-C4H8, and C2H2, J. Phys. Chem. 99 (1995) 3201-3205. Zeep R.G. and D.M. Cline; Rates of Direct Photolysis in Aquatic Environment, Environ. Sci. and Technol. 11 (1997) 359-366.

Kinetic Study of the Temperature Dependence of the OH Initiated Oxidation of Dimethyl Sulphide Mihaela Albu1,2, Ian Barnes1, and Raluca Mocanu 2 1

Bergische Universität Wuppertal, Physikalische Chemie/FB C, Gaußstraße 20, D-42097 Wuppertal, Germany 2 ”Al.I. Cuza” University of Iasi, Faculty of Chemistry, Department of Analytical Chemistry, Carol I Boulevard 11, 700506 Iasi, Romania Key Words: Dimethyl sulphide, Hydroxy radical, Kinetics, Marine atmosphere, Photooxidation

Introduction Dimethyl sulphide (CH3SCH3, DMS) is the dominant natural sulphur compound emitted from the world’s oceans (Berresheim et al., 1995, Urbanski and Wine, 1999), accounting for about one quarter of global sulphur gas emissions. Oceanic DMS, through its oxidation products, is proposed to play a key role in climate regulation, especially in the remote marine atmosphere (Charlson et al., 1987). At the low NOx levels that characterize the remote marine boundary layer, the reaction with OH appears to be the most important loss process for DMS: OH + DMS ĺ products

(1)

The kinetics of the reaction of OH radicals with DMS have been extensively studied by a variety of direct and competitive techniques. The three early studies: two flashphotolysis/resonance fluorescence studies (in Ar/SF6 buffer gases) by Atkinson et al. (1978) and Kurylo (1978) and a competitive rate study (in 760 Torr of air) by Cox and Sheppard (1980) agreed that the rate coefficient is approximately 9 × 10-12 cm3 molecule-1 s-1 at ambient temperature. Atkinson et al. and Kurylo reported a small negative activation energy for the reaction. In addition, Atkinson et al. suggested that the mechanism of the reaction was a hydrogen abstraction. These three studies seemed to indicate that the rate coefficient was well defined, and unaffected by the presence of O2. However, in a later flash-photolysis/resonance fluorescence study Wine et al. (1981) obtained a rate coefficient of 4.3 × 10-12 cm3 molecule-1 s-1 and observed a small positive activation energy for the reaction. The authors suggested that in the previous studies, impurities in the DMS samples used may have affected the results. Also, the results of Cox and Sheppard could have been affected by secondary reactions leading to the loss of DMS. MacLeod et al. (1984) measured the rate coefficient at 373 and 573 K (no air present) and a value of 10.4 × 10-12 cm3 molecule-1 s-1 at room temperature was obtained from their data. They also reported a small positive activation energy. Atkinson et al. (1984) reinvestigated the reaction using a relative rate technique, and reported a rate coefficient of 10.3 × 10-12 cm3 molecule-1 s-1 at room temperature in the presence of 760 Torr of air. This was in reasonable agreement with Cox and Sheppard’s value in air, and the first two studies in the absence of air, but was not in agreement with the value of Wine et al. At this time, the situation was confused. But, numerous later studies: Martin et al. (1985), Wallington et al. (1986), Hynes et al. (1986), Hsu et al. (1987), Barnes et al. (1988), Nielsen et al. (1989), Abbatt et al. (1992), Hynes et al. (1995), Barone et al. (1996), and Williams et al. (2001) were carried out. The value of the activation energy was not satisfactorily resolved by these studies: Martin et al. and Wallington et al. reported small negative values for activation energy, whereas Hynes et al. (1986) and Hsu et al. reported small positive values. 223 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 223–230. © 2006 Springer. Printed in the Netherlands.

224

M. Albu et al.

Several authors have shown that rate coefficient was affected by the presence of O2 (Wallington et al. (1986), Hynes et al. (1986), Barnes et al. (1988)). In their comprehensive work, Hynes et al. (1986) found that the effective rate coefficient for OH + DMS and its deuterated analog (DMS-d6) was dependent on oxygen concentration, increasing as the partial pressure of oxygen was increased. The “oxygen enhancement” was the same for both DMS and its deuterated analog, showing no kinetic isotope but a significant negative temperature dependence. The experimental kinetic observations for reaction of OH radicals with DMS and DMS-d6 are consistent with a two-channel reaction mechanism involving both hydrogen abstraction (O2-independent channel) and reversible adduct formation in competition with adduct reaction with O2 (O2-dependent channel): OH + CH3SCH3 ĺ CH3SCH2 + H2O OH + CH3SCH3 (+M) ļ CH3S(OH)CH3 + (M)

(1a) (1b, 1-b)

CH3S(OH)CH3 + O2 ĺ products

(1c)

OH + CD3SCD3 ĺ CD3SCD2 + HDO

(2a)

OH + CD3SCD3 (+M) ļ CD3S(OH)CD3 + (M)

(2b, 2-b)

CD3S(OH)CD3 + O2 ĺ products

(2c)

In the work of Hynes et al. (1986) the temperature dependence of the rate coefficient for DMS + OH in 760 Torr air, together with temperature dependence of the branching ratio were determined. The authors obtained the following expression for kobs for one atmosphere of air: kobs = {T exp(-234/T) + 8.46 x 10-10 exp(7230/T) + 2.68 x 10-10 exp(7810/T) }/ {1.04 x 1011 T + 88.1 exp(7460/T)} The branching ratio for abstraction being given by: k1a/kobs = 9.6 x 10-12 exp(-234/T) /kobs. For many years, the two-channel mechanism has been widely accepted as the operative mechanism for the OH-initiated oxidation of DMS, and the expressions for the rate coefficient and inferred branching ratio recommended as the best currently available for mechanistic interpretation and atmospheric modelling purposes. A recent re-evaluation of the rate coefficient and the branching ratio has been made by Williams et al. (2001) using the pulsed laser photolysis-pulsed laser induced fluorescence (PLP-PLIF) technique. The effective rate coefficient for the reaction of OH + DMS and OH + DMS-d6 was determined as a function of O2 partial pressure at 600 Torr total pressure in N2/O2 mixtures; the temperature was 240 K for DMS and 240, 261, and 298 K for DMS-d6. This new work shows that at low temperatures the currently recommended expression underestimates both the effective rate coefficient for the reaction and also the branching ratio between addition and abstraction. For example, at 261 K a branching ratio of 3.6 was obtained as opposed to a value of 2.8 based on the work of Hynes et al. (1986). At 240 K the discrepancy increases between a measured value of 7.8 and a value of 3.9 using the extrapolated values from the 1986 work of Hynes et al. (the branching ratio is defined here as (kobs-k1a)/k1a). In addition, at 240 K the expression for kobs in 1 atm air based on the work of Hynes et al. (1986) predicts a value which is a factor of 2 lower then the value measured at

Kinetic Study of the Temperature Dependence of the OH Initiated Oxidation

225

600 Torr. Because of the size of these discrepancies at low temperatures, further kinetic studies are necessary to validate the results of Williams et al. (2001). The preferred value for the rate coefficient for reaction (1a) from the review of Atkinson et al. (2004) is k1a = 1.13 x 10-11 exp(-253/T) cm3 molecule-1 s-1 over the temperature range 240 - 400 K (k1a = 4.8 x 10-12 cm3 molecule-1 s-1 at 298 K). This value is based on the studies of Wine et al. (1981), Hynes et al. (1986), Hsu et al. (1987), Abbatt et al. (1992), and Barone et al. (1996). The preferred value for the rate coefficient for reaction (1b) from the review of Atkinson et al. (2004) is k1b = 1.0 x 10-39 [O2] exp(5820/T)/{1 + 5.0 x 10-30 [O2] exp(6280/T)} cm3 molecule-1 s-1 over the temperature range 240 - 360 K (k1b= 1.7 x 1012 cm3 molecule-1 s-1 at 298 K, 1 bar air). This value is based on the studies of Hynes et al. (1986), and Williams et al. (2001). Experimental The rate coefficients for the gas-phase reaction of OH radicals with dimethyl sulphide were determined using a relative rate method. The kinetic experiments were performed at a total pressure of 1000 mbar diluent gas (N2, synthetic air, or N2/O2 mixtures), at three different O2 partial pressure (~0 mbar, 205 mbar, and 500 mbar) and six different temperatures (250 K, 260 K, 270 K, 280 K, 290 K, and 299 K). The photolysis of hydrogen peroxide (H2O2) was used as the OH radical source: H2O2 + hȞ (Ȝ < 370 nm) ĺ 2 OH Ethene, propene, and iso-butene have been used as the reference organic compounds. The OH radical rate coefficients of these compounds are reliably established (Atkinson et al., 1997). All the experiments were carried out in a 336 l reaction chamber, which is outlined schematicly in Figure 1.

Figure 1.

Schematic diagram of the 336 l reaction chamber.

226

M. Albu et al.

The reactor is composed of 2 borosilicate glass cylinders of length 150 cm and different diameters which are mounted concentrically one within the other. The inner tube has a diameter of 60 cm. Both ends of the cylinders are closed with stainless steel flanges. Elastic KEL-F-seals (GDF Bauart 314) have been used as the sealing material in the entire reactor. The volume of the reactor after deduction of the volume for various fittings is 336 l. The reactor can be temperature controlled to within ±1 in the range 230 to 300 K. Isododecane (Maloterm P1), which is used as the coolant, is circulated continuously through the cavity between the two glass cylinders. The isododecane is cooled by three cryostats (HAAKE N3) connected in series. The temperature of the coolant in the cavity is controlled continuously by 8 platinum resistance thermometers. The reactor is isolated by several layers of the elastomeric nitrile rubber material Armaflex and the entire system is encased in a rectangular aluminium housing. The space between the outer housing and the end flanges is flushed with dry air to avoid condensation of water. Two other quartz glass tubes with diameters of 230 and 244 mm are mounted (concentrically one within the other) and centrally within the inner main reactor cylinder. The smaller of these central tubes contains 12 photolysis lamps. The lamps are of 2 types: 6 Philips TLA 40W (300 ” Ȝ ” 450 nm Ȝmax = 360 nm) and 6 Philips TUV 40W (main emission at Ȝ = 254 nm). The area between the glass cylinders can be cooled either by flowing coolant or dry air. The reactor is equipped with 2 White mirrors systems, one for UV spectroscopy and one for FT-IR spectroscopy. The mirror systems are mounted on the end flanges. The in and out coupling of the probing light is through windows (KCl for the IR and quartz for the UV) mounted on the flange containing the field mirror of the White system. The adjacent flange has connectors for the pumping system, pressure gauges and gas bottles. In addition an easily accessible smaller flange is mounted on the main flange for checking the condition of the reactor interior or placing samples inside the reactor. This small flange also contains a window for visual inspection of samples or chemical gas mixtures during measurements. The spectrometers and the outer reactor housing are connected with one another by a plastic housing, which is flushed with dry air and through which the analysing light beams pass. The starting concentrations (calculated at the working temperature) of DMS and of the reference organic compounds were in the range 1 to 3 ppm and that of H2O2 was approximately 20 ppm. After introducing the reactants into the chamber, the reaction chamber was filled to 1000 mbar with nitrogen, synthetic air, or nitrogen/oxygen mixtures in accordance with the type of experiment which was performed. The concentration-time behaviours of DMS and the reference organic compounds were followed over 30 - 35 min time periods by FT-IR spectroscopy. After acquiring 10 - 15 spectra of the mixture in dark (to control that no dark reaction takes place and to follow possible wall losses), the photooxidation was initiated by switching on the lamps (6 Philips TLA 40W) and then 15 - 20 spectra were recorded. Spectra were obtained by co-adding 64 scans which yielded a time resolution of 1 min. All the reactants were used without further purification. Dimethyl sulphide was supplied by Sigma Aldrich with a stated purity of 99%. H2O2 (85% wt) was supplied by Peroxide Chemie GmbH. The reference hydrocarbons: ethene, propene, and iso-buthene, were supplied by Messer-Griesheim with stated purities of 99.95%, 99.95%, and • 99%, respectively. The following gases were supplied by Messer-Griesheim: nitrogen (99.999%),

Kinetic Study of the Temperature Dependence of the OH Initiated Oxidation

227

oxygen (99.999%), and hydrocarbon free synthetic air (99.999%) with a N2:O2 ratio of 20.5%:79.5%. Results and Discussions The kinetic data obtained in this work shows that, for all the temperatures studied, the rate coefficient for DMS + OH is dependent on the O2 partial pressure, increasing as the O2 partial pressure is increased from ~0 to 500 mbar. At a fixed O2 partial pressure, the rate coefficient is observed to increase on decreasing the temperature from 299 to 250 K; for 205 mbar and 500 mbar O2 partial pressure, the temperature dependence of the rate coefficient was found to be strongly. Figure 2 shows the temperature dependence of the rate coefficient for OH + DMS obtained in this work in 1000 mbar air and at an oxygen partial pressure of 500 mbar. For comparison this figure includes the fit of Hynes et al. (1986) in 1 atm of air and the values calculated for 205 mbar oxygen partial pressure and 500 mbar oxygen partial pressure using the new preferred expressions from the review of Atkinson et al. (2004). From the Figure 2 it can be seen that at 1 atm of air and in the temperature range from 270 to 299 K, the values obtained in this study are in reasonable agreement with those obtained from the fit of Hynes et al. (1986); but at low temperatures the values measured in this work are higher in comparison with those of Hynes et al. (1986). The values obtained here are, however, in reasonable agreement with the new preferred values from the review of Atkinson et al. (2004), which took into account the data from the new work of Williams et al. (2001). 4 Hynes et al.(1986) fit: 760 Torr of air

Atkinson et al. (2004) (pO2 = 205 mbar)

3

This work (pO2 = 500 mbar)

Atkinson et al. (2004) (pO2 = 500 mbar)

2

kDMS [10

-11

3

-1

-1

cm molecule s ]

This work (1000 mbar of air)

1

0 230

240

250

260

270

280

290

300

310

T (K)

Figure 2.

Variation in kDMS as a function of temperature.

320

228

M. Albu et al.

10

Hynes et al.(1986) fit: 760 Torr of air This w ork (205 mbar O2) + Hynes et al. (1986) - abstraction This w ork (500 mbar O2) + Hynes et al. (1986) - abstraction Williams et al.(2001): 150 Torr of O2 at 600 Torr total pressure This w ork (205 mbar O2) + Atkinson et al. (2004) - abstraction This w ork (500 mbar O2) + Atkinson et al. (2004) - abstraction

branching ratio (addition/abstraction)

8

6

4

2

0 230

240

250

260

270

280

290

300

310

320

T (K)

Figure 3.

Temperature dependence of the branching ratio between the addition and abstraction channel.

Figure 3 shows the temperature dependence of the branching ratio between the addition and abstraction channel for DMS + OH obtained using the data from this work, Hynes et al. (1986), Williams et al. (2001) and Atkinson et al. (2004). For the calculation of the branching ratio between addition and abstraction for this work, values for the abstraction rate coefficients were taken, for comparison purposes, both from the expression reported by Hynes et al. (1986) and also from the new recommendations of Atkinson et al. (2004). The comparison in Figure 3 shows that our data are in good agreement with the more recent data of Williams et al. (2001) and that the work of Hynes et al. (1986) underestimated the ratio between addition and abstraction at low temperatures. Conclusions In this work rate coefficients for the reaction OH + DMS ĺ products have been determined over different conditions of temperature (250 K - 299 K) and oxygen partial pressure (~0 mbar - 500 mbar). For all temperature studied, the rate coefficient increased as the partial pressure of oxygen was increased. The kinetic results in the temperature range (270 K - 299 K) are in good agreement with the work of Hynes et al. (1986). The results at low temperature support the more recent observation of Williams et al. (2001) for a large positive roll-off in the rate coefficient for OH + DMS compared with the work of Hynes et al. (1986).

Kinetic Study of the Temperature Dependence of the OH Initiated Oxidation

229

Acknowledgement Financial support for this work within the EL CID 5th Framework Programme Project of the European Commission is gratefully acknowledged. References Abbatt, J.P.D., F.F. Fenter and J.G. Anderson; High-pressure discharge flow kinetics study of OH + CH3SCH3, CH3SSCH3 o products from 297 to 368 K, J. Phys. Chem. 96 (1992), 1780-1785. Atkinson, R., R.A. Perry and J.N. Pitts Jr.; Rate constants for the reaction of OH radicals with COS, CS2 and CH3SCH3 over the temperature range 299-430 K, Chem. Phys. Lett. 54 (1978), 14-18. Atkinson, R., J.N. Pitts Jr. and S.M. Aschmann; Tropospheric reactions of dimethyl sulfide with NO3 and OH radicals, J. Phys. Chem. 88 (1984), 1584-1587. Atkinson, R.; Gas-phase tropospheric chemistry of volatile organic compounds: 1. alkanes and alkenes, J. Phys. Chem. Ref. Data 26 (1997), 215-290. Atkinson, R., D.L. Baulch, R.A. Cox, J.N. Crowley, R.F. Hampson, R.G. Hynes, M.E. Jenkin, M.J. Rossi and J. Troe; Evaluated kinetic and photochemical data for atmospheric chemistry: volume 1 – gas phase reactions of Ox, HOx, NOx and SOx species, Atmos. Chem. Phys. 4 (2004), 1461-1738. Barnes, I., V. Bastian and K.H. Becker; Kinetics and mechanisms of the reaction of OH radicals with dimethyl sulphide, Int. J. Chem. Kinet. 20 (1988), 415-431. Barone, S.B., A.A. Turnipseed and A.R. Ravishankara; Reaction of OH with dimethyl sulfide (DMS). 1. Equilibrium constant for OH + DMS reaction and the kinetics of the OH.DMS + O2 reaction, J. Phys. Chem. 100 (1996), 14694-14702. Berresheim, H., P.H. Wine and D.D. Davis; Sulphur in the atmosphere, in: H.B. Singh (ed.), Composition, Chemistry, and Climate of the Atmosphere, Van Nostrand Reinhold, New York (1995), 251-307. Charlson, R.J., J.E. Lovelock, M.O. Andreae and S.G. Warren; Oceanic phytoplankton, atmospheric sulphur, cloud albedo and climate, Nature 326 (1987), 655-661. Cox, R.A. and D. Sheppard; Reactions of OH radicals with gaseous sulphur compounds, Nature 284 (1980), 330-331. Hsu, Y.-C., D.-S. Chen and Y.-P. Lee; Rate constant for the reaction of OH radicals with dimethyl sulfide, Int. J. Chem. Kinet. 19 (1987), 1073-1082. Hynes, A.J., P.H. Wine and D.H. Semmes; Kinetics and mechanism of OH reactions with organic sulphides, J. Phys. Chem. 90 (1986), 4148-4156. Hynes, A.J., R.B. Stoker, A.J. Pounds, T. McKay, J.D. Bradshaw, J.M. Nicovich and P.H. Wine; A mechanistic study of the reaction of OH with dimethyl-d6 sulfide. Direct observation of adduct formation and the kinetics of the adduct reaction with O2, J. Phys. Chem. 99 (1995), 16967-16975. Kurylo, M.J.; Flash photolysis resonance fluorescence investigation of the reaction of OH radicals with dimethyl sulfide, Chem. Phys. Lett. 58 (1978), 233-237. MacLeod, H., J.L. Jourdain, G. Poulet and G. Le Bras; Kinetic study of reactions of some organic sulfur comounds with OH radicals, Atmos. Environ. 18 (1984), 2621-2626. Martin, D., J.L. Jourdain and G. LeBras; Kinetic study for the reactions of OH radicals with dimethylsulfide, diethylsulfide, tetrahydrothiophene, and thiophene, Int. J. Chem. Kin. 17 (1985), 1247-1261. Nielsen, O.J., H.W. Sidebottom, L. Nelson, J.J. Treacy and D.J. O’Farrell; An absolute and relative rate study of the reaction of OH radicals with dimethyl sulfide, Int. J. Chem. Kin. 21 (1989), 1101-1112. Urbanski, S.P. and P.H. Wine; Chemistry of gas phase organic sulfur-centred radicals, in: Z.B. Alfassi (ed.), SCentred Radicals, John Wiley and Sons (1999), 97-140. Wallington, T.J., R. Atkinson, E.C. Tuazon and S.M. Aschmann; The reaction of OH radicals with dimethyl sulfide, Int. J. Chem. Kin. 18 (1986), 837-846.

230

M. Albu et al.

Williams, M.B., P. Campuzano-Jost, D. Bauer and A.J. Hynes; Kinetic and mechanistic studies of the OHinitiated oxidation of dimethylsulfide at low temperature – a reevaluation of the rate coefficient and branching ratio, Chem. Phys. Lett. 344 (2001), 61-67. Wine, P.H., N.M. Kreutter, C.A. Gump and A.R. Ravishankara; Kinetics of OH reactions with the atmospheric sulfur compounds H2S, CH3SH, CH3SCH3, and CH3SSCH3, J. Phys. Chem. 85 (1981), 2660-2665.

Environmental Chamber Studies of Ozone Formation Potentials of Volatile Organic Compounds William P. L. Carter College of Engineering Center for Environmental Research and Technology, University of California, Riverside, California, 92521, USA Key Words: Environmental chambers, Atmospheric chemical mechanisms, Ozone, Aromatic Hydrocarbons, Petroleum Distillates, Coatings VOCs; SAPRC-99 Mechanism

Abstract Selected results of environmental chamber experiments carried out in the new large indoor environmental chamber at the University of California at Riverside (UCR) that are relevant to quantifying ozone impacts of volatile organic compounds (VOCs) are described. Issues and data needs for quantification of VOC reactivities towards ground-level ozone are described. The ability of the current SAPRC-99 chemical mechanism to simulate recent data from this chamber concerning the ozone formation from irradiations of an ambient reactive organic gas (ROG) - NOx mixture and several new aromatics experiments are discussed. It was found that the mechanism consistently under predicts ozone formation in ambient surrogate - NOx experiments at low ROG/NOx ratios, and also under predicts the effects of adding CO to aromatic - NOx irradiations. The two problems may be related and suggest problems with the formulation of current mechanisms for atmospheric reactions of aromatics. Background Many different types of volatile organic compounds (VOCs) are emitted into the atmosphere, where they can affect photochemical ozone formation and other measures of air quality. Because they can react in the atmospheres at different rates and with different mechanisms, the different types of VOCs can vary significantly in their effects on air quality. The effect of a VOC on ozone formation in a particular environment can be measured by its “incremental reactivity”, which is defined as the amount of additional ozone formed when a small amount of the VOC is added to the environment, divided by the amount added. Although this can be measured in environmental chamber experiments, incremental reactivities in such experiment cannot be assumed to be the same as incremental reactivities in the atmosphere (Carter and Atkinson, 1989; Carter et al., 1995). This is because it is not currently practical to duplicate in an experiment all the environmental factors that affect relative reactivities; and, even if it were, the results would only be applicable to a single type of environment. The only practical means to assess atmospheric reactivity, and how it varies among different environments, is to estimate its atmospheric ozone impacts using airshed models. However, airshed model calculations are no more reliable than the chemical mechanisms upon which they are based. While the initial atmospheric reaction rates for most VOCs are reasonably well known or at least can be estimated, for most VOCs the subsequent reactions of the radicals formed are complex and have uncertainties that can significantly affect predictions of atmospheric impacts. For this reason, environmental chamber experiments and other experimental measurements of reactivity are necessary to test and 231 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 231–240. © 2006 Springer. Printed in the Netherlands.

232

W. P. L. Carter

verify the predictive capabilities of the chemical mechanisms used to calculate atmospheric reactivities. As discussed in a companion paper (Carter, 2004a), under U.S. EPA funding the “UCR EPA” chamber has been constructed at the University of California at Riverside (UCR) in order to address some of the deficiencies of environmental chambers previously used for mechanism evaluation. This chamber has been generating mechanism evaluation data since early 2003, both for general ozone formation mechanism evaluation, and for evaluation of ozone impacts of selected coatings VOCs. The data have been used to evaluate the SAPRC-99 gas-phase mechanism (Carter, 2000), which, together with the European Master Chemical Mechanism (MCM) (Jenkin et al, 1997; Saunders et al, 2003; Jenkin et al, 2003; MCM 2004) represents one of the two state-of-the-art mechanisms that can be used to assess ozone impacts of many individual VOCs. Although the evaluation of the MCM against chamber data is currently underway (e.g., Pilling, 2004; Pinho et al, 2003), SAPRC-99 is still the most extensively evaluated of the mechanisms against available chamber data (Carter, 2000). In this paper, we will discuss the results of the evaluations of the SAPRC-99 mechanism with UCR EPA chamber data relevant to chemical mechanism uncertainty with respect to predictions of effects of VOCs on ozone. The data discussed here are relevant “base mechanism” uncertainty, which refer to general uncertainties in the predictions of ozone formation from reactive organic gas (ROG) and NOx mixtures representing ambient pollution. This is important to prediction of effects of overall VOC and NOx controls and also to appropriately representing the chemistry of base case conditions when assessing incremental reactivities of various types of individual types of VOCs. The specific experiments discussed here are ambient ROG surrogate - NOx experiments at varying ROG and NOx levels, and experiments to evaluate the photooxidation mechanism for aromatics, which are important and highly reactive components of the ambient ROG mixture. Prediction of Ozone formation in Ambient ROG Surrogate - NOx Experiments As discussed in our companion presentation (Carter, 2004a), the initial experiments carried out in the UCR EPA chamber consisted of a large number of ambient ROG surrogate NOx experiments carried out at varying initial NOx and ROG levels. The ambient ROG surrogate composition was derived as discussed by Carter et al (1995) and consisted of a simplified mixture of designed to represent the major classes of hydrocarbons and aldehydes measured in ambient urban atmospheres, with one compound used to represent each model species used in current condensed lumped-molecule mechanisms. The eight representative compounds used were n-butane, n-octane, ethene, propene, trans-2-butene, toluene, m-xylene, and formaldehyde. The initial NOx levels varied from 2 to ~300 ppb and the initial ROG levels varied from ~0.2 to 4.2 ppmC. Figure 1 shows the matrix of initial ROG and NOx levels of the ambient ROG surrogate - NOx experiments carried out in the UCR EPA chamber for mechanism evaluation. The different symbols indicate the project for which the experiment was conducted and whether radical measurement data were made, but for the purpose of this discussion they all are the same experiment except for the varying ROG and NOx levels. Note that this is a loglog plot, with the initial NOx varying by almost two orders of magnitude, and the initial ROG varying by over one order of magnitude. The solid line show the ROG/NOx levels that the SAPRC-99 mechanism predicts give the highest ozone levels and the dotted line shows the

Environmental Chamber Studies of Ozone Formation Potentials

233

ROG/NOx levels where the mechanism predicts O3 has the greatest sensitive to changes in VOC levels (MIR line). It can be seen that the matrix includes experiments in both the VOC sensitive and NOx-sensitive regions as defined by these lines.

Initial NOx (ppb)

1000

Initial Evaluation Runs

VOC Sensitive Region

OBM Runs without Radical Data

100

OBM Runs with OH, HO2 Data Used for Base Case for Reactivity Runs

10

Calculated Maximum O3 Line

NOx-Limited Region

Calculated MIR Line

1 0.1

1

10

Initial ROG (ppmC)

Figure 1.

Matrix of initial ROG and NOx levels in the ambient ROG surrogate - NOx experiments carried out in the UCR EPA chamber. The initial ROG and NOx levels used in the base case experiments for the incremental reactivity studies of the coatings VOC emissions are also shown.

The two large open circles on Figure 1 show the ROG and NOx levels chosen to serve as the “base case” in the incremental reactivity experiments that are currently underway in this chamber to assess ozone impacts of various different types of VOCs (Carter, 2004c). The 2530 ppb NOx levels were chosen to be representative of pollution episodes of interest in California, based on input provided by the staff of the California Air Resources Board, which is funding most of the current reactivity studies. The ROG/NOx ratios were chosen to represent two sets of conditions of NOx availability relevant to VOC reactivity. The lower ROG/NOx ratio was chosen to represent the relatively higher NOx conditions of maximum incremental reactivity (MIR) where ozone is most sensitive to VOCs, to approximate the conditions used to derive the widely-used MIR ozone reactivity scale (Carter, 1994). The higher ROG/NOx ratio was chosen to be one-half that yielding maximum ozone levels, and is used to represent conditions where ozone is NOx-limited, but not so NOx-limited that VOC reactivity is irrelevant. These are referred to in the subsequent discussion as the “MIR” and “MOIR/2” base cases, respectively. The ability of the SAPRC-99 mechanism to simulate ozone formation in representative MIR and MOIR/2 base case experiments is shown on Figure 2. It can be seen that good reproducibility in ozone formation is observed in the replicate experiments, but that the model consistently under predicts O3 formation in the MIR experiments, though it gives much better predictions of ozone in the experiment at the higher ROG/NOx ratio. This tendency to under

234

W. P. L. Carter

MIR Base Case

MOIR/2 Base Case

(NOx = 30 ppb, ROG = 0.5 ppmC)

(NOx = 25 ppb, ROG = 1 ppmC)

0.16

0.16

0.12

0.12

Model

0.08

EPA110B

EPA114A

0.04

Ozone (ppm)

Ozone (ppm)

predict O3 in the base case experiment representing MIR conditions presents a problem when using incremental reactivity MIR experiments for evaluating mechanism of added VOCs.

Model

0.08

EPA095B

0.04

EPA124B

EPA123A

EPA127B

EPA126A 0.00

0.00 0

60

120

180

240

Time (minutes)

Figure 2.

300

360

0

60

120

180

240

300

360

Time (minutes)

Experimental and calculated concentration-time plots for ozone representative base case experiments used for VOC reactivity assessment.

in

The dependence of the tendency for the SAPRC-99 mechanism under prediction bias on the initial ROG/NOx ratio is shown on Figure 3. In order to provide useful mechanism evaluations at the lowest ROG/NOx ratios where O3 formation is suppressed by NO, the model predictions of ozone formed + NO oxidized, measured by ([O3]final-[NO]final) ([O3]initial-[NO]initial) or '([O3]-[NO]), is shown. Comparable results of a recent SAPRC-99 mechanism evaluation using ambient surrogate - NOx data from other low NOx environmental chamber datasets (Carter, 2004b, and references therein) are also shown. In order to place the ROG/NOx ratios for the various chamber datasets on the same basis, the biases are plotted against ROG/NOx ratios normalized to the ratio calculated to give the highest ozone formations for the conditions of the various chambers. Although this was not evident from the previous mechanism evaluation using higher NOx chamber data with higher chamber effects (Carter, 2000; see also Carter, 2004b), the results of this work show a consistent tendency of increase in ozone under prediction bias with decreasing ROG/NOx ratios for the ambient surrogate experiments in the UCR EPA chamber. This bias may also be present in the simulations of the ambient surrogate experiments in the TVA chamber, which were also carried out at relatively low NOx levels (Simonaitis and Bailey, 1995; Bailey et al, 1996; Carter, 2004b). The SAPRC-99 mechanism under predicts O3 formation and NO oxidation in almost all the ambient surrogate experiments carried out in the Australian CSIRO chamber that were used for low NOx mechanism evaluation (Hess et al, 1992; Johnson et al, 1997; Carter, 2004b). Although this might be due to characterization uncertainties with this outdoor chamber (Carter, 2004b), if the relative ROG/NOx ratios are taken into account the results are entirely consistent with the trends observed when modelling the UCR EPA experiments.

Environmental Chamber Studies of Ozone Formation Potentials

235

Plots of the bias for the SAPRC-99 mechanism to under predict NO oxidation and O3 formation against normalized ROG/NOx ratios for ambient surrogate NOx experiments in three environmental chambers.

'(O3-NO) Model underprediction Bias

Figure 3.

80%

SAPRC-99 (Range shows effect of varying HONO offgasing rates)

MOIR

60% SAPRC-99 with Reduced k(OH + NO2)

40%

Carbon Bond 4 Mechanism

20% 0% Approx. MIR

-20% 0.1

Figure 4.

1

10

Plots of the mechanism under prediction bias for model simulations UCR EPA ambient surrogate - NOx experiments against normalized ROG/NOx ratios for the SAPRC-99 and Carbon Bond mechanisms. The SAPRC-99 calculations show effects of varying the HONO offgasing chamber effect parameters and using a lower OH + NO2 rate constant.

Figure 4 shows the effects of using different mechanisms or mechanism parameters on plots of the under prediction bias the UCR EPA ambient surrogate experiments. The filled diamonds show the same results as shown on Figure 3 for this chamber, and the error bars show the effects of varying the HONO offgasing chamber effect parameter from zero to the

236

W. P. L. Carter

maximum level indicated by the variability of the characterization experiments. This is probably the most important chamber effects parameter that would affect the results of model simulations of these experiments, since the radical source caused by the HONO photolysis will affect the results of the low ROG/NOx experiments while the NOx input caused by this process will affect the high ROG/NOx runs. While this variable chamber effect has a nonnegligible effect, it is not enough to explain the under prediction bias observed at the low ROG/NOx ratios. On the other hand, it could account for the apparent over prediction bias at the highest ROG/NOx ratios. Various sensitivity studies have indicated that uncertainties in the rate constant for the reaction of OH with NO2, an important radical termination and NOx process, could have effects of model simulations of O3. In order to assess whether this could account for the bias observed, simulations were carried out with the rate constant reduced by ~20%, which is probably reflective of its current uncertainty. It can be seen that this reduces the bias somewhat, but does not eliminate it. Figure 4 also shows that the low ROG/NOx model under prediction bias is significantly worse for the Carbon Bond 4 mechanism (Gery et al, 1989). Although somewhat out-of-date, this mechanism is still widely used in many models in the United States because of its computational efficiency. The reason for this under prediction bias is still uncertain, and it is being investigated by conducting experiments with modified ROG surrogates and as part of an ongoing project to update the SAPRC mechanism. We suspect that is probably due to problems with the current mechanism for aromatics, which are highly reactive components of the ambient ROG surrogate whose mechanism remain highly uncertain. Additional evidence for problems with the current aromatics mechanisms is discussed in the following section. New Aromatics Mechanism Evaluation Results As discussed in our companion paper (Carter, 2004a), the initial evaluation experiments in the UCR EPA chamber included experiments with the representative aromatic hydrocarbons toluene and m-xylene. Aromatics are important components of ambient ROG mixtures because of their high reactivity, and both toluene and m-xylene are present in the ambient ROG surrogate used in the experiments discussed in the previous section. Aromatics have the most uncertain mechanisms of all the compounds present in this surrogate. Although there have been extensive studies of the atmospheric reactions of aromatics and their reactive products (e.g., see Calvert et al, 2002, and references therein), available information is still insufficient to derive predictive mechanisms. The greatest uncertainty concerns the highly reactive but incompletely characterized aromatic ring-opening products, whose subsequent reactions appear to be significant radical initiators, since ignoring their reactions results in significant under predictions of aromatic reactivity. The current SAPRC-99 mechanism and represents the reactions in a highly simplified manner, using a few parameterized model species whose yields and photolysis rates were adjusted to optimize fits of model simulations to aromatic - NOx chamber experiments (Carter, 2000). The results of the toluene - NOx and m-xylene - NOx experiments in the UCR EPA were reasonably consistent with the predictions of the current SAPRC-99 mechanism, though the mechanism had a slight bias towards over predicting O3 in the experiments (see Carter, 2004a for a summary of the fits to total NO oxidation and O3 formation in the experiments).

Environmental Chamber Studies of Ozone Formation Potentials

237

Examples of experimental and calculated ozone in these runs are shown in Figure 5, where the open squares are the experimental data for representative aromatics - NOx experiments, and the dotted lines are the model simulations of those runs. A reasonably good fit of the model to these experiments is expected because the parameters in the mechanisms for those compounds were adjusted to optimize model simulations of such experiments. However, the NOx levels in these experiments were considerably lower than those in the runs used to develop the SAPRC-99 mechanisms, and there was a concern that the parameterization may not extrapolate to lower NOx levels. Apparently the parameterization employed is no less applicable for NOx levels as low as ~5 ppb than it is for the higher NOx levels in the experiments used in the mechanism development. Toluene

m-Xylene

Concentration (ppm)

Ozone Change Ozone (effect of CO) EPA077 (~25 ppb NOx)

Ozone Change (effect of CO) EPA067 (~20 ppb NOx)

Ozone

0.12

0.25

0.06

0.12 0.20

0.09 0.04

0.08

0.15

0.06 0.10

0.02

0.04

0.03

0.05

0.00

0.00 0

120

240

0.00

0

360

120

240

360

Time (minutes)

0.00

0

120

240

360

0

120

240

360

Time (minutes)

Concentration (ppm)

EPA066 (~5 ppb NOx) 0.10

0.04

0.08

0.03

0.06 0.04

Aromatic - NOx experiment

0.02

Aromatic - CO - NOx experiment

0.01

Aromatic - NOx model

0.02

0.00 0.00

0 0

180

360

540

180

360

540

Aromatic - CO - NOx model

-0.01

Time (minutes)

Figure 5.

Experimental and calculated effects of CO addition on selected aromatics NOx experiments carried out in the UCR EPA chamber.

However, the new UCR EPA chamber dataset also includes experiments where the effect of adding CO to aromatics - NOx irradiations was determined. These experiments were carried out because model calculations indicated that the addition of CO would cause a significant enhancement in O3 formation, due to the NO to NO2 conversions caused when CO reacts with the radicals produced in the aromatic photooxidation reactions, and we wanted to test this prediction. In this regard, CO acts as a “radical amplifier”, enhancing the effects of radicals on ozone formation. CO addition is also useful because CO has the simplest possible mechanism to represent other VOCs present in ambient mixtures; its reactions only cause NO to NO2 conversions and its reactions result in formation of no other products or direct radical sources or sinks. Therefore, added CO experiments should provide a test of an aspect of the aromatics mechanisms that is applicable to its effects in ambient simulations, and that is different than the tests provided by aromatic NOx experiments in the absence of other VOCs.

238

W. P. L. Carter

The results of the aromatics - NOx experiments with the added CO were consistent with the model in that the added CO did indeed significantly increase ozone formation, but the amount of O3 increase was about twice what the model predicted. This is shown on Figure 5, where the filled diamonds show the O3 in the aromatic - NOx experiments or the change in O3 caused by adding the CO, and the solid lines show the corresponding model predictions. About the same under prediction of the CO addition effect is seen for m-xylene as toluene, but the under prediction may be greater in the lower NOx toluene experiment. The under prediction of the effect of the added CO suggests that the mechanism is under predicting the radical input in the aromatic photooxidation process. The fact that the same mechanism can simulate O3 formation in the aromatics - NOx experiments without the CO indicates that there probably are compensating errors in the mechanism. Further evidence for this comes from the fact that the mechanism overpredicts NO to NO2 conversions in “direct reactivity” aromatic + HONO experiments that are much more sensitive to NO to NO2 conversions in the mechanisms than radical sources (Carter and Malkina, 2002), suggesting that there may be too many NO to NO2 conversions in the aromatics mechanisms. However, reducing the NO to NO2 conversions in the SAPRC-99 aromatics mechanism to fit the direct reactivity data and re-adjusting the radical input parameters to simulate the aromatics - NOx experiments does not fully correct this problem. This is shown on Figure 6, which shows the model predictions of effects of CO, direct reactivity, and O3 formation in toluene - NOx runs with the standard and modified SAPRC-99 toluene mechanisms. Although the modified mechanism predicts an increase in O3 caused by the added CO compared to the standard mechanism, the increase is insufficient. Even increasing the radical input in the model to levels that result in over prediction of O3 in the toluene - NOx experiments is still not sufficient for the model to simulate the full effect of the CO addition. Effect of CO

Direct Reactivity

Toluene - NOx Runs

(EPA074B vs EPA074A)

0.05

0.00 0

120 240 Time (minutes)

360

0.3 0.4

0.2 0.1

0.2

0.0 2

4 6 Toluene Added (ppm)

Data

Standard Mechanism Re-Adjusted

8

'([O3]-[NO]) (ppm)

0.10

CTC079

0.4

0.6

'([O3]-[NO]) Change (ppm)

'([O3]-[NO]) Change (ppm)

0.15

0.0 0

120

240

360

CTC026

0.6 0.4 0.2

Radical Yield Increased Further 0.0 0

120

240

360

Time (minutes)

Figure 6.

Effects of adjustments to the SAPRC-99 toluene photooxidation mechanism on simulations of representative toluene experiments experiments.

Environmental Chamber Studies of Ozone Formation Potentials

239

Thus the new data suggest that there is a fundamental problem with the formulation of the current aromatics mechanism, and more extensive modifications are required in order for model simulations to be consistent with the available data. We have been funded by the California Air Resources Board to update the SAPRC mechanism, and developing an improved aromatics mechanism is a priority in this project. Work on a new, more explicit mechanism that incorporates more recent laboratory data such as that summarized by Calvert et al (2004) is now underway, but progress is slow. Presently, we are encountering problems similar to those discussed by Pilling (2004) for the MCM. It is likely that there are reactions occurring that none of the mechanisms are adequately representing. Conclusions Although the UCR EPA chamber has only been in operation for a relatively short time, it has already obtained useful information concerning the performance of current mechanism in predicting the effects of VOCs and NOx on ozone formation. As discussed here and also in our companion paper (Carter, 2004a), the SAPRC-99 mechanism predicts O3 formation reasonably well in low NOx experiments, and in ambient simulation experiments at the high ROG/NOx levels where maximum ozone formation potentials are achieved. However, the new data indicate problems with the mechanisms that were not previously realized. The SAPRC99 mechanism consistently under predicts O3 formation in the lower ROG/NOx experiments where O3 formation is most sensitive to VOCs, and the problem is even worse for CB4. Other experiments indicate that there are problems with the formulation with the current aromatics photooxidation mechanisms. It is possible that the problems with the under prediction at low ROG/NOx ratios may be caused by problems with the aromatics mechanisms. Experimental and mechanism development work to investigate and hopefully resolve these problems is underway References Bailey, E. M., C. H. Copeland and R. Simonaitis; Smog chamber studies at low VOC and NOx concentrations,” Report on Interagency Agreement DW64936024 to EPA/NREL, Research Triangle Park, NC. (1996). Calvert, J. G., R. Atkinson, K. K. Becker, R. M. Kamens, J. H. Seinfeld, T. J. Wallington, and G. Yarwood; The mechanisms of atmospheric oxidation of aromatic hydrocarbons, New York: Oxford University Press, 556 pgs. (2002) Carter, W. P. L.; Development of ozone reactivity scales for volatile organic compounds,” J. Air & Waste Manage. Assoc., 44, 881-899. (1994) Carter, W. P. L.; Documentation of the SAPRC-99 chemical mechanism for VOC reactivity assessment,” Report to the California Air Resources Board, Contracts 92-329 and 95-308, May 8 (2000). Available at http://www.cert.ucr.edu/~carter/absts.htm#saprc99. Carter, W. P. L.; The UCR EPA Environmental Chamber. Presented at the NATO EST-ARW Workshop on Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, Zakopane, Poland, October 1-4 (2004a). Carter, W. P. L.; Evaluation of a gas-phase atmospheric reaction mechanism for low NOx conditions,” Final Report to California Air Resources Board Contract No. 01-305, May 5 (2004b). Available at http://www.cert.ucr.edu/~carter/absts.htm#lnoxrptf Carter, W. P. L.: Main information page for “Evaluation of atmospheric ozone impacts of selected coatings VOC emissions” and “Environmental chamber studies of VOC species in architectural coatings and mobile source emissions”. http://www.cert.ucr.edu/~carter/coatings. Last modified April 17. (2004c)

240

W. P. L. Carter

Carter, W. P. L. and R. Atkinson; A computer modeling study of incremental hydrocarbon reactivity, Environ. Sci. Technol., 23, 864 (1989) Carter, W. P. L., D. Luo, I. L. Malkina, and J. A. Pierce; Environmental chamber studies of atmospheric reactivities of volatile organic compounds. effects of varying ROG surrogate and NOx,” Final report to Coordinating Research Council, Inc., Project ME-9, California Air Resources Board, Contract A0320692, and South Coast Air Quality Management District, Contract C91323. March 24 (1995). Available at http://www.cert.ucr.edu/~carter/absts.htm#rct2rept. Carter, W. P. L. and I. L. Malkina; Development and application of improved methods for measurement of ozone formation potentials of volatile organic compounds, Final report to California Air Resources Board Contract 97-314, May 22 (2002). Available at http://www.cert.ucr.edu/~carter/absts.htm #rmethrpt. Hess G. D., F. Carnovale. M. E., Cope. and G. M. Johnson; The evaluation of some photochemical smog reaction mechanisms – 1 temperature and initial composition effects, Atmos. Environ., 26A, 625-641. (1992) Jenkin, M. E., S. M. Saunders and M. J. Pilling; The tropospheric degradation of volatile organic compounds : a protocol for mechanism development,” Atmos. Environ. 31, 81-104 (1997). Jenkin, M.E., S.M. Saunders, V. Wagner and M.J. Pilling; Protocol for the development of the master chemical mechanism MCMv3 (Part B): Tropospheric degradation of aromatic volatile organic compounds, Atmospheric Chemistry and Physics Discussions, 2, p1905-1938 (2002). Atmospheric Chemistry and Physics, 3, 181-193 (2003) Johnson G., P. C. Nancarrow, A.Quintanar, and M. Azzi; Smog chamber data for testing chemical mechanisms at low VOC to NOX ratio conditions, Final Report to EPA Cooperative Agreement CR-821420, Atmospheric Research and Exposure Assessment Laboratory, U.S. EPA. (1997) MCM; The Master Chemical Mechanism website, http://chmlin9.leeds.ac.uk/MCM/. Undated. Last accessed 10/2004 (2004) Pihno, P. G.; C. A. Pio, and M. E. Jenkin; Evaluation of isoprene degradation on the detailed tropospheric chemical mechanism, MCM v3, using environmental chamber data. Presented at the NATO EST-ARW Workshop on Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, Zakopane, Poland, October 1-4 (2004). Pilling, M. J.; Aromatic hydrocarbon oxidation: the contribution of chamber photooxidation studies. Presented at the NATO EST-ARW Workshop on Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, Zakopane, Poland, October 1-4 (2004). Saunders, S.M., M.E. Jenkin, R.G. Derwent and M.J. Pilling; Protocol for the development of the master chemical mechanism MCMv3 (Part A): Tropospheric degradation of non-aromatic volatile organic compounds, Atmospheric Chemistry and Physics Discussions, 2, p1847-1903 (2002). Atmospheric Chemistry and Physics, 3, 161-180 (2003) Simonaitis, R. and E. M. Bailey; Smog chamber studies at low VOC and NOx concentrations: Phase I,” Report on Interagency Agreement DW64936024 to EPA/NREL, Research Triangle Park, NC. (1995).

Evaluation of the Detailed Tropospheric Chemical Mechanism, MCM v3, Using Environmental Chamber Data: Butane and Its Degradation Products P. G. Pinho1, C.A. Pio1, and M. E. Jenkin2 2

1 Departamento de Ambiente, Escola Superior de Tecnologia de Viseu, 3504-510 Viseu, Portugal Department of Environmental Science and Technology, Imperial Collage London, Silwood Park, Ascot, Berkshire SL5 7PY, UK

Key Words: Butane, Environmental Chamber, Mechanisms, Modelling, Ozone, Troposphere

Introduction The chemistry of ground level atmospheric ozone formation is highly complex and nonlinear and has many uncertainties. As a result no chemical model can be relied upon to give accurate predictions unless it has been evaluated by comparison with experimental data. There are essentially two ways in which a photochemical oxidant model can be evaluated. One compares the predictions of the complete model against field data taken during a historic ozone pollution episode. Another approach relies on the evaluation of individual pollutant compounds, separately. In the case of gas-phase chemical mechanisms, this means evaluating the predictions of the mechanism against results of environmental chamber experiments (Carter et al., 1995a). There are a large number of condensed chemical reaction mechanisms to describe and represent atmospheric chemical transformations involving air pollutants. On the contrary only a very limited number of mechanisms exist with a detailed description of chemical transformation processes for primary and secondary atmospheric species. The Master Chemical Mechanism (MCM) (Jenkin et al., 1997) and SAPRC (Carter and Lurmann, 1991) are two well known examples. The MCM is a near-explicit chemical mechanism that describes the detailed degradation of a series of emitted VOC, and the resultant generation of ozone and other secondary pollutants, under conditions appropriate to the planetary boundary layer. Version 3 of the Master Chemical Mechanism (MCM v3) considers the oxidation of 125 VOC. The complete mechanism comprises 12,691 reactions of 4,351 organic species and 46 associated inorganic reactions, which were defined on the basis of the MCM scheme writing protocols (Jenkin et al., 1997, Saunders et al., 2003; Jenkin et al., 2003). Although MCM v3 has recently been superseded by v 3.1, the chemistry for non-aromatic VOC remains unchanged. Although the entire MCM has been tested in atmospheric models, and through intercomparison with the results of chamber-validated mechanisms (e.g. Derwent et al., 1998; Jenkin et al., 2002), it has only been partially tested using environmental chamber data. It has been used in a number of studies involving the European Photoreactor (EUPHORE) in Valencia (EUPHORE, 2002), providing the basis for validation of the mechanisms for selected VOC. MCM v3 chemistry has thus already been tested for the photo-oxidation of Dpinene-NOx mixtures at comparatively low NOx concentrations (Saunders et al., 2003). The aromatic mechanisms in MCM v3 and MCM v3.1 have also been evaluated against a set of smog-chamber experiments; the evaluation was focused on four representative species of the 241 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 241–252. © 2006 Springer. Printed in the Netherlands.

P. G. Pinho et al.

242

class of aromatic compounds: benzene, toluene, p-xylene and 1,3,5-TMB (Bloss et al., 2004a; 2004b). The main objective of the present study is the evaluation and refinement of MCM v3, using the large environmental chamber dataset of the Statewide Air Pollution Research Center (SAPRC) at the University of California (Carter et al., 1995a). Initial evaluation of the representation of chamber-dependent processes has been carried out using butane-NOX photooxidation data and the MCM v3 degradation chemistry for butane, and the results of this evaluation are presented. Within this assessment, photooxidation experiments involving the major carbonyl products formed during butane degradation, namely methylethyl ketone (MEK), acetaldehyde (CH3CHO) and formaldehyde (HCHO) were also considered. Throughout the study, the performance of the MCM v3 chemistry was simultaneously compared with that of the corresponding chemistry in the SAPRC-99 mechanism, which was developed and optimized in conjunction with the chamber datasets (Carter, 2000). This work therefore develops a base for the evaluation of degradation mechanism for a number of VOC represented in MCM v3, using the SAPRC environmental chamber datasets. Chamber data The mechanistic evaluation was carried out using the database of the indoor chambers of the Statewide Air Pollution Research Center (SAPRC) at the University of California (Carter et al., 1995a). A set of computer data files containing the SAPRC environmental chamber data was obtained on the Internet by anonymous FTP at carterpc.ucr.edu. Data from the following environmental chambers were used in the present evaluation: Indoor Teflon Chamber #1 (ITC); Indoor Teflon Chamber #2 (ETC); Dividable Teflon Chamber (DTC); Xenon Arc Teflon Chamber (XTC) and CE-CERT Xenon Arc Teflon Chamber (CTC). The available datasets from these chambers were used previously in the development and evaluation of the SAPRC-99 mechanism (Carter, 2000). A large set of chamber experimental runs were employed in the present evaluation. These consisted of 46 butane-NOX-air, 6 MEK-NOX-air, 11 CH3CHO-NOX-air and 24 HCHO-NOX-air experiments, in the initial evaluation of the chamber dependent processes and butane degradation chemistry. The ranges of reagent concentrations for the considered experiments are presented in Table 1. Full details of the experimental conditions and reagent concentrations are available elsewhere (Carter et al., 1995a, Carter, 2000), which includes detailed documentation of the measurement uncertainties associated with the analytical methods employed for the reactants and products, and the precision and accuracy of other experimental parameters (e.g. temperature; light intensity and spectrum of the photolysing radiation). Chamber-dependent processes Wall effects Heterogeneous processes occurring at the chamber walls are probably the most serious problem in evaluation and interpretation of experiments carried out in environmental chambers. These wall effects are known to be non-negligible in all current-generation chambers, and can dominate the results of certain types of experiments (e.g., Carter and Lurmann, 1991, Carter, 2000).

Evaluation of the Detailed Tropospheric Chemical Mechanism, MCM v3 Table 1.

243

Ranges of reagent concentrations for the considered experiments. Butane-NOx

MEK-NOx

CH3CHO-NOx

HCHO-NOx

[VOC] (ppm)

1.47–4.95

7.83–9.49

0.44–1.67

0.29–1.01

[NOx] (ppm)

0.09-0.66

0.09–0.29

0.14–0.28

0.16-0.54

[VOC]/[NOx]

5.22-39.89

1.71–11.73

1.00–3.40

26.61– 99.44

There are three major groups of processes that contribute to chamber wall effects: the introduction of free radicals through wall reactions, the adsorption or desorption of oxidised nitrogen species (NOy), and off-gassing of organics that can lead to OH to HO2 conversion and therefore supplement ozone production in the system (Killus and Whitten, 1990). The chamber wall effects can be characterized for a given chamber, through experiments which are designed to be especially sensitive to the wall effects, but for which the gas phase chemistry is reasonably well understood. Such characterization experiments can be used to create a model for chamber wall effects and chamber-dependent parameters (Carter et al., 1995a), which is usually termed the ‘auxiliary mechanism’ (Jeffries et al., 1992). A database for mechanism evaluation is therefore not complete without recommendations for appropriate inputs to auxiliary mechanisms. Such chamber effects models have already been developed for the SAPRC chambers as part of the process of evaluating SAPRC-99 and related mechanisms (e.g. Carter, 2000), and these form the basis of the auxiliary mechanisms applied in the present study. Photolysis processes The rates of photolysis processes depend on the intensity and spectral distribution of the light source, which vary from one chamber to another. Spectral distributions (based on spectrometer measurements) and absolute light intensities (based on NO2 actinometry experiments) are documented for all the SAPRC chambers (e.g. Carter et al., 1995a, Carter et al., 1995b). The rates of photolysis processes for application in MCM v3 were thus determined using this spectral information in conjunction with absorption cross sections and quantum yield data previously used in conjunction with the MCM, based on the sources summarised by Jenkin et al., (1997). Furthermore, the rates of a number of photolysis reactions were updated during the course of the present study, as discussed further below. Butane photo-oxidation and auxiliary mechanism assessment Chemistry of butane photo-oxidation The complete degradation chemistry of butane, as represented in MCM v3, consists of 538 reactions of 178 species. It can be viewed and downloaded using the subset mechanism assembling facility, available as part of the MCM website. The methodology of mechanism construction has been described in detail by Jenkin et al., 1997 and Saunders et al., 2003. The main features of the degradation chemistry when NOX is present are summarized in Figure 1.

P. G. Pinho et al.

244

butane OH O2

C2 H 5 O

NO

NO

O2

O

O2

HO2

HO2

O2

O

CH3CHO O

OH

NO

O2

O2

OH

MEK

O

NO2

CH3C(O)O2

- NO2

NO

O2

CH3C(O)O2NO2

PAN

CH3C(O)O O2 NO

CH3O

O2

HO2

OH

HCHO

O2

CO2

O2

OH

CO

HO2

HO2

Figure 1.

Schematic representation of the major radical propagation pathways of the OHinitiated degradation of butane, also illustrating the sequential formation of the intermediate products MEK, CH3CHO, HCHO and CO. Reaction steps involving NO are reactions with organic peroxy radical intermediates, which lead to NO-to-NO2 conversion and O3 production (see discussion in text). The reactions of OH with butane and MEK possess additional channels which are fully represented in the mechanism, but not shown in the diagram.

Mechanism evaluation As with the chamber characterisation studies of Carter (Carter, 2000), the series of butane-NOX photo-oxidation experiments was used for initial assessment of the auxiliary mechanism parameters. This system is believed to provide a good test for the chamber wall effects because the degradation chemistry of butane is quite well characterized (Carter and Lurmann, 1991, Carter et al., 1995a), and because of the large set of experiments for which data are available. Initial simulations were carried out to compare the SAPRC-99 and MCM v3 butane mechanisms for the conditions of the butane-NOX experiments, using the auxiliary mechanism parameters defined by Carter, (2000), and a consistent set of inorganic reaction

Evaluation of the Detailed Tropospheric Chemical Mechanism, MCM v3

245

rate coefficients. The MCM v3 and SAPRC-99 butane mechanisms were found to perform very similarly, suggesting a good level of consistency in the understanding and representation of butane oxidation in these mechanisms. To allow a full evaluation, however, it is necessary to consider not only the experimental data for the butane-NOx system, but also the NOx photo-oxidation runs for intermediate products, MEK, CH3CHO and HCHO. The aim of the mechanism refinements for butane degradation, therefore, was to provide an adequate, self-consistent description of all these systems, whilst simultaneously identifying any required modifications to the auxiliary mechanism parameters. Mechanism testing and refinement was therefore carried out by an iterative procedure, considering butane photo-oxidation and each of the sub-systems in turn. The adapted version of MCM v3, resulting from identified modifications, is denoted ‘MCM v3a’ throughout this article. In conjunction with these tests, the opportunity was taken to revise and update the inorganic chemistry parameters. The most notable change from the reference simulations, as performed for SAPRC-99 by Carter, (2000), was a change in the assigned rate coefficient for the reaction of OH with NO2 from a value (at 298 K and 760 Torr) of 8.98 × 10-12 cm3 molecule-1 s-1 (based on JPL, 1997) to a value of 1.05 × 10-11 cm3 molecule-1 s-1 (based on JPL, 2003). This reaction represents an important radical sink in the system, which was found to have a particularly significant impact under the conditions of some of the butane-NOx photo-oxidation experiments. As described further below, this required small compensating increases to be applied to the chamber radical source parameter. As with previous assessments (e.g. Carter, 2000), the quantity D(O3-NO) was used as the main criterion of model performance. This quantity is defined as: D(O3-NO)t = [O3]t [NO]t - ([O3]0 - [NO]0), where [O3]0, [NO]0, and [O3]t, [NO]t are the concentrations of O3 and NO at the beginning of the run, and at time ‘t’, respectively. As described in detail previously (e.g., Carter and Lurmann, 1991, Carter et al., 1995a, Carter, 2000), D(O3-NO) is an indicator of the ability of the mechanism to simulate the chemical processes that cause O3 formation, giving a useful measure, even when O3 is suppressed by the presence of excess NO. In addition, use of this measure allows a direct comparison with the SAPRC-99 published results. The precursor decay rate and formation of the carbonyl products and PAN were also used as criteria of model performance. HCHO-NOX experiments The MCM v3 mechanism was found to under-predict D(O3-NO) in many (but not all) of the runs, principally for the CTC and DTC chambers. This discrepancy was improved through the implementation of revised absorption cross section and quantum yield data for the photolysis of HCHO. The photolysis parameters were updated in line with the latest IUPAC recommendations (Data Sheet P1 updated in 16th May 2002), the preferred cross sections being based on the data of Meller and Moortgat, (2000), and the quantum yields based on data from Smith et al. (2002). The new cross sections are 5-10% higher than the values previously recommended by IUPAC (and adopted for the MCM by Jenkin et al., 1997). The improved performance of the modified mechanism, MCM v3a, in simulating the HCHO-NOX chamber experiments is summarised in Figure 2 and Figure 3. The improvement is particularly apparent for the large set of CTC and DTC chambers runs, which represent the majority of the dataset for HCHO.

246

P. G. Pinho et al.

0,45

0,30

DTC149B

0,60

CTC024

CTC016

0,40

0,25

0,50

0,20

0,40

0,15

0,30

0,10

0,20

0,05

0,10

0,35 0,30 0,25 0,20 0,15 0,10 0,05

0,00

0,00 0

60

120

180

240

300

0,00

0

360

SAPRC 99

60

120

180

240

300

360

0

60

120

180

240

300

360

MCM v3a

Experimental

MCM v3

Plots of experimental and calculated D(O3-NO) (ppm) vs. time (min) for the HCHO – NOx -air experiments.

Figure 2. 0,50

0,40

DTC149B

0,45

0,90

CTC024

0,35

0,80

0,30

0,70

0,40 0,35 0,30

0,25

0,25

0,20

0,20

0,15

CTC016

0,60 0,50 0,40 0,30

0,15

0,10

0,20

0,10 0,05

0,05

0,00

0,00 0

60

120

180

240

300

0,10 0,00

0

360

60

120

180

240

MCM v3

SAPRC 99

300

360

0

60

120

Experimental

180

240

300

360

MCM v3a

Plots of experimental and calculated HCHO (ppm) vs. time (min) for the HCHO – NOx -air experiments.

Figure 3.

0,25

0,40

DTC387B

0,40

CTC014

0,35

0,35

0,30

0,30

0,25

0,25

0,20

0,20

0,15

0,15

0,10

0,10

0,05

0,05

DTC150B

0,20

0,15

0,10

0,05

0,00

0,00

0,00 0

60

120

180

240

SAPRC 99

Figure 4.

300

360

0

60

MCM v3

120

180

240

300

360

Experimental

0

60

120

180

240

300

360

MCM v3a

Plots of experimental and calculated D(O3-NO) (ppm) vs. time (min) for the CH3CHO – NOx –air experiments.

CH3CHO-NOX experiments Apart from a small secondary influence of the changes in the HCHO photolysis parameters, no further modifications were required to the acetaldehyde mechanism. As shown in Figure 4 to Figure 7, MCM v3a provides a very good description of D(O3-NO) in the complete set of CH3CHO-NOX chamber experiments.

Evaluation of the Detailed Tropospheric Chemical Mechanism, MCM v3 0,6

1,0

DTC387B

0,5

1,8

CTC014

0,9

1,6

0,8

1,4

0,7

0,4

247 DTC150B

1,2

0,6

1,0

0,5

0,3

0,8

0,4 0,2

0,6

0,3

0,4

0,2

0,1

0,2

0,1 0,0

0,0 0

60

120

180

240

300

0,0

0

360

SAPRC 99

60

120

180

240

300

360

0

0,050

180

240

300

360

MCM v3a

Experimental

MCM v3

0,10

DTC387B

0,08

CTC014

0,045

0,09

0,040

0,08

0,035

0,07

0,030

0,06

0,05

0,025

0,05

0,04

0,020

0,04

0,015

0,03

0,010

0,02

0,005

0,01

0,000

0,00 60

120

Plots of experimental and calculated CH3CHO (ppm) vs. time (min) for the CH3CHO – NOx –air experiments.

Figure 5.

0

60

120

180

240

300

0,06

0,03 0,02 0,01 0,00

0

360

DTC150B

0,07

60

120

180

240

360

Experimental

MCM v3

SAPRC 99

300

0

60

120

180

240

300

360

MCM v3a

Plots of experimental and calculated HCHO (ppm) vs. time (min) for the CH3CHO – NOx –air experiments.

Figure 6.

0,012

0,05

DTC387B

0,010

CTC014

0,04

0,008 0,03

0,006 0,02

SAPRC 99 MCM v3 Experimental MCM v3a

0,004 0,01

0,002 0,000

0,00

0

60

Figure 7.

120

180

240

300

360

0

60

120

180

240

300

360

Plots of experimental and calculated PAN (ppm) vs. time (min) for the CH3CHO – NOx –air experiments.

MEK-NOX experiments The MCM v3 mechanism was found to over-predict D(O3-NO) in many of the chamber runs, particularly in the early stages of the experiment. This discrepancy was improved by reducing the quantum yield for the photolysis of MEK (reaction (7)). In MCM v3, a wavelength-independent value of 0.34 is applied, based on the atmospheric pressure data from Raber and Moortgat, (1996). That study demonstrated that the quantum yield decreases with increasing pressure, but also quoted increasingly wide uncertainty limits on the quantum

248

P. G. Pinho et al.

yield at pressures approaching one atmosphere. The value was therefore optimized within MCM v3a, to give the best fit to the MEK-NOx chamber data (see Figure 8 to Figure 12). This led to a best fit value of 0.17, which is close to the extremity of the uncertainty limit of the determination of Raber and Moortgat, (1996). This value is also consistent with the value of 0.15 obtained by Carter, 2000 during evaluation and optimization of the SAPRC-99 mechanism. 0,8

0,6

DTC337A

0,7

0,5

CTC178A

0,5

DTC361A

0,4

0,6

0,4

0,5

0,3

0,3

0,4

0,2

0,3

0,2

0,2

0,1

0,1

0,1

0,0

0,0 0

60

120

180

240

300

0,0

0

360

60

SAPRC 99

120

180

240

300

360

0

60

120

Experimental

MCM v3

180

240

300

360

MCM v3a

Plots of experimental and calculated D(O3-NO) (ppm) vs. time (min) for the MEK – NOx –air experiments.

Figure 8.

8,0

9,2

DTC337A

7,8

9,3

CTC178A

7,6

DTC361A

9,2

9,1

9,1

9,0

7,4

9,0

8,9

7,2

8,9

8,8

7,0

8,8

6,8

8,7

6,6

8,6

0

60

120

180

240

300

360

8,7 8,6

0

SAPRC 99

60

120

180

240

MCM v3

300

360

0

60

120

180

240

300

360

MCM v3a

Experimental

Plots of experimental and calculated MEK (ppm) vs. time (min) for the MEK – NOx –air experiments.

Figure 9.

0,10

0,12

DTC337A

0,08

CTC178A

0,09 0,08

DTC361A

0,07

0,10

0,06

0,07

0,08

0,05

0,06 0,05

0,06

0,04

0,04

0,03

0,04

0,03

0,02

0,02

0,02

0,01

0,01 0,00

0,00 0

60

120

180

240

SAPRC 99

Figure 10.

300

360

0,00

0

60

MCM v3

120

180

240

300

360

Experimental

0

60

120

180

240

300

360

MCM v3a

Plots of experimental and calculated HCHO (ppm) vs. time (min) for the MEK – NOx –air experiments.

Evaluation of the Detailed Tropospheric Chemical Mechanism, MCM v3 0,18

0,25

DTC337A

249

DTC361A

0,16

0,20

0,14 0,12

0,15

0,10 0,08

0,10

SAPRC 99 MCM v3 Experimental MCM v3a

0,06 0,04

0,05

0,02

0,00

0,00 0

60

120

180

240

300

0

360

60

120

180

240

300

360

Plots of experimental and calculated CH3CHO (ppm) vs. time (min) for the MEK – NOx –air experiments.

Figure 11.

0,16

0,14

DTC337A

0,14

0,09

CTC178A

DTC361A

0,08

0,12

0,07

0,12

0,10

0,06

0,10 0,08

0,05

0,06

0,04

0,08 0,06

0,03 0,04

0,04

0,02 0,02

0,02 0,00

0,01 0,00

0,00

0

60

120

180

240

300

360

0

SAPRC 99

Figure 12.

60

MCM v3

120

180

240

300

360

Experimental

0

60

120

180

240

300

360

MCM v3a

Plots of experimental and calculated PAN (ppm) vs. time (min) for the MEK – NOx –air experiments. The experimental PAN data in the CTC chamber runs appear anomalous and are not well described by either MCM v3a or SAPRC99.

Butane-NOX experiments Because there is no primary radical source during butane photo-oxidation, this parent system was the most sensitive to changes in the chamber radical source parameter. In conjunction with the optimizations described above, it was found that the best fit to the butane-NOX photo-oxidation data was obtained by multiplying the Carter chamber radical source values used in the SAPRC evaluation (Carter, 2000) by a factor of 1.35 for the ITC and ETC chambers, and by a factor of 1.2 for the DTC, CTC and XTC chambers. With these changes, the adapted MCM v3a butane mechanism was found to give generally good fits to D(O3-NO) data in the complete series of butane–NOX photo-oxidation experiments. The scatter in the results was indicative of run-to-run variability, with most of the data being fit by the mechanism to within +/- 30%, with no consistent biases. Examples of D(O3-NO), butane decay rate and formation of CH3CHO and MEK are presented in Figure 13 to Figure 16. As indicated above, the butane, HCHO, CH3CHO and MEK systems were considered iteratively, such that the optimized chamber radical sources and the changes made to the photolysis parameters for HCHO and MEK led to a self-consistent description of all the systems.

250

P. G. Pinho et al.

0,12

0,25

ITC939

0,10

0,30

ITC533

ITC507

0,25

0,20

0,20

0,08

0,15

0,15

0,06

0,10

0,10

0,04

0,05

0,02

0,05 0,00

0,00

0,00 0

60

120

180

240

300

0

360

SAPRC 99

60

120

180

240

300

0

360

60

120

Experimental

MCM v3

180

240

300

360

MCM v3a

Plots of experimental and calculated D(O3-NO) (ppm) vs. time (min) for the butane – NOx -air experiments.

Figure 13.

4,90

3,00

ITC939

3,80

ITC533

ITC507

4,85 2,95

3,75

2,90

3,70

2,85

3,65

4,80

4,75

4,70

4,65

2,80

0

60

120

180

240

300

360

3,60

0

SAPRC 99

60

120

180

240

MCM v3

300

360

0

60

120

180

240

300

360

MCM v3a

Experimental

Plots of experimental and calculated butane (ppm) vs. time (min) for the butane – NOx -air experiments.

Figure 14.

0,040

0,06

ITC939

0,035

0,07

ITC533

0,030

0,05

0,04

0,025

ITC507

0,06

0,05

0,04 0,020

0,03

0,03 0,015

0,02

0,02

0,010 0,01

0,005 0,000

0,01 0,00

0,00

0

60

120

180

240

SAPRC 99

Figure 15.

300

360

0

60

MCM v3

120

180

240

300

360

Experimental

0

60

120

180

240

300

360

MCM v3a

Plots of experimental and calculated CH3CHO (ppm) vs. time (min) for the butane – NOx -air experiments.

Evaluation of the Detailed Tropospheric Chemical Mechanism, MCM v3 0,030

0,045

ITC939

0,08

ITC533

0,040

ITC507

0,07

0,025

0,035

0,06

0,030

0,020

251

0,05

0,025 0,015

0,04

0,020 0,03

0,015

0,010

0,010

0,02

0,005

0,01

0,005

0,000

0,000 0

60

120

180

240

300

SAPRC 99

Figure 16 .

360

0,00

0

60

MCM v3

120

180

240

300

360

Experimental

0

60

120

180

240

300

360

MCM v3a

Plots of experimental and calculated MEK (ppm) vs. time (min) for the butane – NOx -air experiments.

Summary and Conclusions The MCM v3 butane degradation chemistry has been evaluated using chamber data on the photo-oxidation of butane, and on photo-oxidation of its degradation products, methylethyl ketone (MEK), acetaldehyde (CH3CHO) and formaldehyde (HCHO), in conjunction with an initial evaluation of the chamber-dependent auxiliary mechanisms for the series of relevant chambers. The MCM v3 mechanism for butane was found to provide an acceptable reaction framework for describing the NOx-photo-oxidation experiments on the above systems, although a number of parameter modifications and refinements were identified and introduced which resulted in an improved performance. These generally relate to the magnitude of sources of free radicals from carbonyl photolysis processes, which are currently under review in MCM development activities. Specifically, recommendations are made to update the photolysis parameters for HCHO, MEK, either in line with data reported since the MCM mechanism development protocol (Jenkin et al., 1997), or on the basis of optimization in the current study. Throughout the study, the performance of MCM v3 and the refined mechanism (denoted MCM v3a) was simultaneously compared with that of the SAPRC-99 mechanism, which was developed and optimized in conjunction with the chamber datasets (Carter, 2000). The performance of MCM v3a was found to be very similar to that of SAPRC-99. References Bloss C., V. Wagner, A. Bonzanini, M. E. Jenkin, K. Wirtz, M. Martin-Reviejo and M. J. Pilling; Evaluation of detailed aromatic mechanisms (MCM v3 and MCM v3.1) against environmental chamber data, Atmospheric Chemistry and Physics Discussions, 4 (2004a) 5683-5731. Bloss C., V. Wagner, M.E. Jenkin, R. Volkamer, W. J. Bloss, J. D. Lee, D. E. Heard, K. Wirtz, M. MartinReviejo, G. Rea, J. C. Wenger and M. J. Pilling; Development of a detailed chemical mechanism (MCMv3.1) for the atmospheric oxidation of aromatic hydrocarbons, Atmospheric Chemistry and Physics Discussions, 4 (2004b) 5733-5788. Carter, W. P. L. and F. W. Lurmann; Evaluation of a Detailed Gas-Phase Atmospheric Reaction Mechanism using Environmental Chamber Data, Atmospheric Environment 25A (1991) 2771-2806. Carter W. P. L., D. Luo, I. L. Malkina and D. Fitz,; The University of California, Riverside Environmental Chamber Data Base for Evaluating Oxidant Mechanisms: Indoor Chamber Experiments Through 1993, EPA/600/SR-96/078, U.S. Environmental Protection Agency, Research Triangle Park, NC (1995a).

252

P. G. Pinho et al.

Carter W. P. L., D. Luo, I. L. Malkina and J. A. Pierce; Environmental Chamber Studies of Atmospheric Reactivities of Volatile Organic Compounds. Effects of Varying Chamber and Light Source, Final report to National Renewable Energy Laboratory, Contract XZ-2-12075, Coordinating Research Council, Inc., Project M-9, California Air Resources Board, Contract A032-0692, and South Coast Air Quality Management District, Contract C91323 (1995b). Carter W. P. L.; Documentation of the SAPRC-99 Chemical Mechanism for VOC reactivity Assessment, Final Report to California Air Resources Board Contract 92-329 and Contract 95-308, Air Pollution Research Center and College of Engineering Center for Environmental Research and Technology University of California Riverside, California (2000). Derwent, R.G., M.E. Jenkin, S.M. Saunders and M.J. Pilling; Photochemical ozone creation potentials for organic compounds in North West Europe calculated with a master chemical mechanism, Atmospheric Environment, 32 (1998) 2429-2441. IUPAC Subcommittee on Gas Kinetic Data Evaluation, http://www.iupac-kinetic.ch.cam.ac.uk/. Jeffries H.E., M.W. Gery and W.P.L. Carter; Protocols for evaluating oxidant mechanisms for urban and regional models, EPA 600/R-92/112. Atmospheric Research and Exposure Assessment Laboratory, ORD, USEPA, Research Triangle Park, North Carolina (1992). Jenkin M.E., S.M. Saunders and M.J. Pilling; The tropospheric degradation of volatile organic compounds: A protocol for mechanism development, Atmospheric Environment, 31 (1997) 81-104. Jenkin M.E., T.J. Davies and J.R. Stedman; The origin and day-of-week dependence of photochemical episodes in the UK. Atmospheric Environment, 36 (2002) 999-1012. Jenkin, M. E., S. M. Saunders, V. Wagner and M. J. Pilling; Protocol for the development of the Master Chemical Mechanism, MCM v3, Part B: tropospheric degradation of aromatic volatile organic compounds. Atmospheric Chemistry and Physics, 3 (2003) 181-193. Killus, J. P. and G. Z. Whitten; Backgroud reactivity in smog chambers. Int. J. Chem. Kinet. 22 (1990) 547-575. Meller R. E. and G. K. Moortgat; Temperature dependence of the absorption cross sections of formaldehyde between 223 and 323 K in the wavelength range 225-375 nm, J. Geophys.Res. 105 (2000) 7089-7101. Raber W.H. and G.K. Moortgat; Progress and Problems in Atmospheric Chemistry, edited by J. Barker, 318-373, World Scientific Publ. Co., Singapore (1996). Saunders S.M., M.E. Jenkin, R.G. Derwent and M.J. Pilling; Protocol for the development of the Master Chemical Mechanism, MCM v3, Part A: tropospheric degradation of non-aromatic volatile organic compounds, Atmospheric Chemistry and Physics, 3 (2003) 161-180. Smith G. D., L. T. Molina and M. Molina; Measurement of Radical Quantum Yields from Formaldehyde Photolysis between 269 and 339 nm, J.Phys.Chem. A, 106 (2002) 1233-1240.

Studies on Nitrate-Affected SO2 Oxidation and Their Perspectives Wanda Pasiuk-Bronikowska and Tadeusz Bronikowski Institute of Physical Chemistry, Polish Academy of Sciences, Kasprzaka 44/52, 01-224 Warsaw, Poland Key Words: Autoxidation mechanism, Cross-activation, Nitric radicals, Reactors, Sulphoxy radicals

Introduction Atmospheric chemistry has not yet been fully understood. One of the main reasons is the complex interaction between components of the gas phase and of the condensed phases. In this contribution we would like to consider, by example of the reaction between aqueous SO2 and molecular oxygen, occurring in the presence of nitrate, some aspects of such an interaction which may be of interest both to the researchers and modellers of cloud chemistry. Sources of nitrate in tropospheric droplets Let us first consider briefly the sources of nitrate in the troposphere aqueous phase. In modelling tropospheric multiphase chemistry, a number of more or less recognized gas phase and aqueous phase reactions are included which produce nitric species. We focus on the following main modes leading to nitrate within a droplet: Mode I. Dissolution of nitric species generated in a gas phase;

NO2 + OH + M ĺ HNO3 + M during daytime (Crutzen, 1995)

NO3 +

ĺ HNO3 +

The reaction between nitric radicals and pinonaldehyde (Waengberg et al., 1997) is only one example of the role of organic compounds in the gas phase formation of nitric acid in the atmosphere. At this point the heterogeneous reactions between gaseous species and solid or humidified NaCl should also be mentioned as a source of easily water soluble NaNO3: NO3 + NaCls ĺ NaNO3 + Cl ClONO2 + NaCls ĺ NaNO3 + Cl2 Mode II. Uptake of trace N-compounds, accompanied by a chemical reaction in a droplet: 2NO2 + H2O ĺ NO2Ǧ + NO3Ǧ + 2H+ N2O5 + H2O ĺ 2NO3Ǧ + 2H+ HNO4 + HSO3Ǧ ĺ NO3Ǧ + SO42Ǧ + 2H+ NO3 + org ĺ NO3Ǧ + org• 253 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 253–260. © 2006 Springer. Printed in the Netherlands.

254

W. Pasiuk-Bronikowska and T. Bronikowski

Reactions with water remove reactive nitrogen oxides from the gas phase and, moreover, transform nitrogen oxides into reservoir species, nitric acid and its anions. Reactions involving very active radicals generated in the gas phase, like NO3, are less probable in a droplet bulk. Because of the high activity of these radicals, their gas-phase concentrations and mass fluxes are relatively low, so that the radicals should rather become exhausted on the gas-liquid interface. The mechanism of such reactions given by Herrmann et al. (1996) for aromatic hydrocarbons clearly shows the various pathways leading to nitric acid. Mode III. Uptake of nitric acid during the condensation of water. The results obtained so far in field experiments (Laj et al., 1997) do not allow to establish whether nitrate molecules become incorporated within droplets by nucleation at the onset of cloud or fog or if they are scavenged by the droplets from the interstitial air. Nitrate anions are one of the major ionic species present in the Po Valley fog water drops (Decesari et al., 2001). Recently, the Black Sea was shown to release NO3Ǧ rich aerosols into the troposphere due to the bubble bursting process (Karakas et al., 2001). In general, aerosol nitrate is estimated to double in the next half century (Thiemens et al., 2003). Nitric acid is one of the so called “reservoir species” of the atmosphere (Barker, 1995). This means that one may expect the relatively easy release of temporarily “stored” nitric radicals when such a reservoir species is attacked by other radicals. Among these radicals are sulphate radical anions, SO4• Ǧ, formed as intermediates in the autoxidation of S(IV), in the oxidation of aqueous SO2 by molecular oxygen. We shall discuss consequences of this fact in the course of this contribution. It is why sulphur chemistry in the presence of nitrates becomes an area that needs to be further developed, as the cross-activation of these trace components may be important for understanding some field observations. Nitrate as an inhibitor of SO2 autoxidation to sulphuric acid The autoxidation of S(IV) is known to occur by a long chain mechanism (Table 1). Such a mechanism comprises three sequences: initiation, propagation and termination. It is essential to understand clearly that the initiation and termination steps have rates lower, by at least of an order of magnitude, than those of propagation steps. The intervention of nitrate anions in the mechanism of S(IV) autoxidation depends on the concentration of these anions. At relatively low nitrate concentrations, the rate of scavenging SO4• Ǧ radical anions attains low values, comparable with the rate of chain termination. In this instance two specific cases may be distinguished (Table 2): Ɣ The radical product of the reaction between nitrate and SO4• Ǧ (NO3) attains too low concentrations to influence the overall reaction kinetics. This is the classic case of inhibition due to a reaction in which reactive radicals are replaced by other ones, which are inactive under the conditions of the experiments. Ɣ The accumulation of NO3 radicals becomes essential, so that the activity of these radicals is noticeable. The reacting system is switched to the case when the secondary initiation emerges, thus reducing the effect of the nitrate inhibitor. The net effect is the retardation of the rate of S(IV) autoxidation.

Studies on Nitrate-Affected SO2 Oxidation and Their Perspectives

255

Classic mechanism of a long chain reaction.

Table 1. initiation:

-

I

-

MIII + SO32 / HSO 3

-

-

MII + SO3 /( SO3 + H+)

propagation: SO3 + O2 SO5 kp1 SO5 + SO32 /HSO 3 SO52 /HSO5 + 2 SO5 + SO3 /HSO 3 SO42 /HSO4 + 2 SO4 + SO3 /HSO 3 SO42 /HSO4 + nonradical S(IV) oxidation: SO52 + SO32 /HSO 3 2SO 42 /(2SO 42

- - - - termination: 2SO S O -+ O

-

I

I I I

I

I

5

2

2 8

2

-

ki

-

SO 3 (rate controlling step) SO4 kp3 SO 3 kp4

-+ H

+

)

kp2

kp5

(nonradical products )

kt

I

Interference of nitrates (low concentration).

Table 2.

Termination (classic): 2SO5 S2O82 + O2

I

-

-

kt

Termination (nitrate scavenger): NO3 + SO4 NO3 + SO42 ksc (radical product not effective kinetically)

I

-

-

-

Retardation (nitrate scavenger): NO3 + SO4 NO3 + SO42 (radical product kinetically effective)

I

-

-

Initiation (primary ): MIII + HSO3 MII + SO3 + H+ (secondary): NO3 + HSO3 NO3 + SO3 + H+

I I

-

-

-

-

-

k sc

ki,p ki,s

Two plots are shown in Figure 1, the upper one indicates the concentration of S(VI) (the S(IV) autoxidation product) growing with the time of reaction performed under quasisteady state conditions and the lower one which gives the rate - scavenger (nitrate) concentration profile. It is evident that the rate of S(IV) autoxidation is a non-monotonous function of the concentration of nitrate. The question arises, what is the influence of nitrate on the rate of S(IV) autoxidation at still higher concentrations of this compound.

W. Pasiuk-Bronikowska and T. Bronikowski

256

' [ SVI ]

In this light the question arises what will happen at still higher concentrations of nitrate; should one expect the total compensation of the inhibition by nitrate as observed at low concentration, when the nitrate concentration attains a sufficiently high value? 1

3

Figure 1. Effect of nitrates (low concentration region) on the rate of S(IV) autoxidation: accumulation of the reaction product, S(VI), in time (upper plots); rate – scavenger (nitrate) concentration profile.

2

time

r

?

z1 3

z2

z

Cases of the intervention of nitrate: 1. No nitrate in the reacting system; 2. Nitrate present at very low concentration (no effect of nitric radicals on chain initiation); 3. Nitrate present at low concentration (substantial effect of nitric radicals on chain initiation).

> Sc@ Nitrate as a catalyst of SO2 autoxidation to sulphuric acid Sea aerosols initially have an ionic composition of sea water but quickly loose water to attain equilibrium with water vapour in the surrounding air, simultaneously cumulating reactive trace gases and undergoing chemical reactions (Pszenny et al., 1998). Also cloud droplets may undergo dehydration, depending on air humidity. The concentration of nitrate may reach very high values, a mole per dm3 and over, when droplets evaporate almost to dryness. High concentrations of nitrate should also be expected in the case when initially dry or humidified solid NaCl aerosols are reacted with nitric radicals, NO3, or chlorine nitrate, ClONO2, reactions recently discussed, for instance, by Gershenzon et al. (2002). Our recent study on the effect of nitrate at high concentrations led to the results which give ground for inferring that the intervention of nitrates may be also at the level of chain propagation (Pasiuk-Bronikowska et al., 2004). Then, the rate of S(IV) autoxidation substantially exceeds that attained in the absence of nitrate. This is illustrated in Figure 2. When the concentration of nitrate attains a sufficiently high level, the reaction rate, r, starts to grow above the value for the uninhibited reaction. Such an acceleration of S(VI) formation in the presence of high amounts of nitrate indicates that the latter compound must react in the rate determining step. As in this part of our experiments the excess of nitrate with respect to S(IV) was immense, it seemed reasonable to explain the observed changes in the S(IV) autoxidation kinetics assuming that:

Studies on Nitrate-Affected SO2 Oxidation and Their Perspectives

257

Ɣ the presence of sulphate radicals in the reacting system is negligible, Ɣ the hydrogen sulphite anions in the rate controlling step, suitable in the case of low concentrations of nitrate, are replaced by nitrate anions.

Effect of nitrates (concentration region extended) on the rate of S(IV) autoxidation.

Figure 2.

As shown in Table 3, nitrate anions consumed in the rate controlling step are reformed in the next step involving nitric radicals and S(IV). This is a sort of a catalytic cycle, where Interference of nitrates (high concentration).

Table 3. initiation:

-

propagation: SO3 + O2

I

MIII + HSO3–

-

MII + •SO3 – + H+

k’i

SO k (rate controlling step) SO - + NO SO - + NO NO + HSO - NO - + SO - + H k nonradical S(IV) oxidation: SO - + HSO 2SO - + H k I I I

•–

5

p1

–

5

3

– 3

2– 5

– 3

I

3

5

– 3

•

2– 4

3

+

3

+

p

p5

termination: 2SO5 S 2O82 + O2 (nonradical products )

I

-

-

I

2–

kt

W. Pasiuk-Bronikowska and T. Bronikowski

258

nitrate anions play a role of a catalyst. The location of point 4 (white spot in Figure 2) is evidence for the nitrate catalysis. These surprising results shed light on the possible important role of nitrate in the atmospheric autoxidation of SO2 by molecular oxygen. In a certain sense they coincide with the conclusions derived by Leriche et al. (2000) who reported on a multiphase box model for a remote region of the troposphere applied to a set of field measurements in France. It is commonly believed that the main oxidant at low pH is H2O2. To account for the observed conversion of S(IV) to S(VI) at very low concentrations of H2O2 Leriche et al. postulated the decisive role of peroxy nitric acid, HNO4. These authors included in their model of Nchemistry in droplets the following reaction: NO2 + HO2/O2Ǧ ĺ HNO4/ NO4Ǧ According to Amels et al. (1996), the S(IV) oxidation by peroxy nitric acid can proceed very rapidly with the simultaneous formation of nitrate: HNO4 + HSO3Ǧ ĺ NO3Ǧ + SO42Ǧ + 2H+ Both oxidation pathways, the NO3Ǧ catalysed oxidation of S(IV) by molecular oxygen and the oxidation of S(IV) by peroxy nitric acid lead to the formation of sulphate anions and further studies are required to decide, which of the two pathways is the mot important (or perhaps both) under real atmospheric conditions. A chamber reactor as a tool for scaling up A variety of reactor types are known which may be roughly divided into two categories: homogeneous and heterogeneous ones, depending on whether the studied reaction involves reactants present in a single phase or whether they must be drawn from two or more phases. The autoxidation of SO2 is an instructive example of a reaction which leads to different kinetic observations depending on the way it investigated in a particular reactor. In the case of a single-phase laboratory reactor, shown schematically in Table 4, the reactor volume is totally occupied by liquid. The autoxidation of SO2 occurs between reactants, including oxygen, pre-dissolved in the liquid. Thus, the only reaction rate limitations come from the properties of a measuring device (Clark-type electrode) and from the exhaustion of reactants. In a two-phase reactor, with liquid as a continuous phase, the reaction may be observed for a longer time, as the gas reactant is continuously replenished. Here, the intrinsic reaction rate limitation becomes the rate of oxygen supply across the oxygen/water interface (including that of gas bubbles). In a three-phase reactor, also with liquid as a continuous phase, the oxygen/water (slurry) interface is planar, what allows the determination of the rate of oxygen diffusion. A solid suspension is a source of the reacting solute supplied by dissolution. Usually, in a chamber reactor (see the bottom row in Table 4) the complex chemical systems existing in the atmosphere gas phase are approximated. Especially advantageous for studying gas reactions are the two characteristic features of a chamber reactor: a relatively large volume and rather insignificant wall effects. Examples of effectively studied tropospheric reactions in chamber reactors are: photo-dissociation and oxidation of a selection of organic compounds, the latter reactions with such oxidants as ozone, OH and NO3. In many

Studies on Nitrate-Affected SO2 Oxidation and Their Perspectives

259

instances, the studies were focused on gas-phase reaction mechanisms and, to a lesser extent, on the gas-phase reaction kinetics. With the introduction of water droplets or wet aerosol particles into the interior of a chamber, mass transfer in the sense of microphysics emerges as a coordinate part of the problem of the proper approximation of chemical transformations in such a system. To quantitatively describe the mass transfer/chemical reaction interplay is not a trivial task. It requires a set of data on gas-side and liquid-side diffusion to be collected and included in existing models.

Applicability of some typical reactors.

Table 4.

TYPE

REACTOR SCHEME oxygen sensor

liquid-phase (single-phase reactor)

oxygen inlet

gas-liquid (two-phase reactor)

gas-liquid-solid (multiphase reactor)

pH sens or or oxygen probe

air inle t outle t conductometric probe

solid suspension

gas-phase

instrume ntation instrumentation

gas-liquid gas-liquid-solid

instrumentation

REACTOR CHARACTERISTICS • volume: 4x10-3 dm3; • reaction time: several sec; • gas reactant predissolved; • formation of secondary products avoided; • serves to determine kinetic laws and rate constants; • volume: 0.10 dm3; • reaction time: tens of min; • gas reactant supplied by diffusion; • formation of secondary products avoided; • serves to determine kinetic laws and rate constants;

• volume: 0.15 dm3; • reaction time: tens of min; • gas reactant supplied by diffusion; • solute reactant supplied by solid dissolution; • formation of secondary products avoided; • serves to determine kinetic laws and rate constants;

• volume: up to ca. 2x10 5 dm3 (200 m3); • reaction time: hours to days; • formation of secondary products allowed; • serves the determination of chemical composition changes in time (product identification and distribution); • influence of mass transfer (convection/diffusion)

?

We would like to raise the question of the inconveniency resulting from the complexity of cloud chemistry models, at least when taking into account the required computing power and time. To simplify the model approach we suggest treating a cloud not as an aqueous phase volume, but as a system of mixed phases, working as a whole. It would require confining a cloud in a chamber reactor and determining new kinetic parameters as if it were a somewhat modified gas-phase. Conclusions The widely held view that reactions of relatively low rate constants may be safely neglected in the modelling of aqueous phase chemistry should be verified. As shown for the cross-activation of S(IV) and nitrate, the neglected reactions may accelerate, possibly by orders of magnitude, due to an increase of the concentration of inorganic components in desiccating aerosols.

260

W. Pasiuk-Bronikowska and T. Bronikowski

The use of a chamber reactor for “confining a cloud in a cage” and treating it as a modified gas phase is suggested, here, as an approach to reasonably simplified modelling of chemical composition changes in real clouds. A comparison between the results obtained using a large volume chamber reactor and those calculated from the aqueous phase model, based on data collected using small laboratory reactors, should shed light on the importance of gas-side and liquid-side resistances to mass transfer. Acknowledgements This work, being the contribution to the ESF programme INTROP, is a logical continuation of our study developed within the frames of he EUREKA project, EUROTRAC2. The financial support from the Scientific Committee for Research in Poland is gratefully acknowledged. References Amels, P., H. Elias, U. Goetz, U. Steingens and K. J. Wannowius; Kinetic investigation of the stability of peroxonitric acid and of its reaction with sulphur(IV) in aqueous solution, in: P. Warneck (ed.), Heterogeneous and Liquid-Phase Processes, Springer, Berlin (1996) 77-88. Barker, J. R.; A brief introduction to atmospheric chemistry, in: J. R. Barker (ed.), Progress and Problems in Atmospheric Chemistry, World Scientific, London (1995) 1-33. Crutzen, P. J.; Overview of tropospheric chemistry: developments during the past quater century and a look ahead, Faraday Discuss., 100 (1995) 1-21. Decesari, S., M. C. Facchini, E. Matta, M. Mircea, S. Fuzzi and J.-P. Putaud; Organic and inorganic solutes in atmospheric aerosol: a full characterisation approach, in: P.M.Midgley, M.J.Reuter, M.Williams (eds), Proc. EUROTRAC-2 Symposium 2000 “Transport and Chemical Transformation in the Troposphere”, Springer, Berlin (2001) 274-277. Gershenzon, M. Yu., V. M. Grigorieva, S. D. Il’in, R. G. Remorov, D. V. Shestakov, V. V. Zelenov, E. A .Aparina and M. Yu. Gershenzon; Heterogeneous reactions affecting chlorine activation in the troposphere, in: I. Barnes (ed.), Global Atmospheric Change and its Impact on Regional Air Quality, Kluwer Academic Publishers, Dordrecht (2002) 109-113. Herrmann, H., H.-W. Jacobi, G. Raabe, A. Reese, Th. Umschlag and R. Zellner; Free radical reactions in the tropospheric aqueous phase, in: B. Larsen, B. Versino, G. Angeletti, (eds), The Oxidizing Capacity of the Troposphere, Report EUR 17482 (1997) 503-512. Laj, P., S. Fuzzi, L. Ricci, A. Berner, D. Schell, M. Wendisch, W. Wobrock, G. Frank and B. Martinssonn; Chemical inhomogeneities in fog droplets during the 1994 CHEMDROP experiment, in: B. Larsen, B. Versino, G. Angeletti (eds), The Oxidizing Capacity of the Troposphere, Report EUR 17482 (1997) 513521. Karakas, D. and S. G. Tuncel; Long-range transport of aerosols to Turkish western Black Sea region, in: P.M.Midgley, M.J.Reuter, M.Williams (eds), Proc. EUROTRAC-2 Symposium 2000 “Transport and Chemical Transformation in the Troposphere”, Springer, Berlin (2001) 290-293. Leriche, M., D. Voisin, N. Chaumerliac, A. Monod and B. Aumont; A model for tropospheric multiphase chemistry: application to one cloudy event during the CIME experiment, Atmos. Environ., 34 (2000) 5015-5036. Pasiuk-Bronikowska, W., T. Bronikowski and M. Ulejczyk; Intervention of inorganic radical scavengers during the autoxidation of S(IV), Environ. Sci. Technol., submitted for publication. Pszenny, A., W. Keene, C. O’Dowd and M. Smith; Sea salt aerosols, tropospheric sulfur cycling, and climate forcing, IGACtivities Newsletter (11) (1998) 6-12. Thiemens, M., G. M. Michalski, A. Romero and J. R. McCabe; Mass independent oxygen and sulfur isotopic compositions of environmental sulfate and nitrate. A new probe of atmospheric, hydrospheric and geological processes, Geophys. Res. Abstracts 5 (2003) no. 04359. Waengberg, I., I. Barnes and K.-H. Becker; NO3 radical initiated oxidation of Į-pinene, in: B. Larsen, B. Versino, G. Angeletti (eds), The Oxidizing Capacity of the Troposphere, Report EUR 17482 (1997) 104-108.

Heterogeneous and Aqueous-Phase Transformations of Isoprene Krzysztof J. Rudzinski Institute of Physical Chemistry of the PAS, Kasprzaka 44/52, 01-224 Warsaw, Poland Key Words: Isoprene, Troposphere, Feedback, Sulphate Radical, Oxidation

Introduction Isoprene is a conjugated diene (2-methyl-buta-1,3-diene), volatile and hardly soluble in water; under normal pressure it boils at 34 oC (Merck, 1999) and dissolves up to 1.47×10-2 M at 21.5 oC, with a Henry’s constant of 0.027 mole kg-1 atm-1 at 25 oC (NIST, 2001). Isoprene is a metabolite in plants, microbes, animals and humans, and a major biogenic trace compound emitted to the atmosphere. It is very reactive towards atmospheric gas-phase oxidants such as hydroxyl and nitrate radicals or ozone. At higher concentrations, 220 – 7000 ppm, it is carcinogenic to rodents and possibly carcinogenic to humans (Melnick and Sills, 2001). Terrestrial vegetation produces most of the emitted isoprene – 500 Tg C per year (Guenther et al., 1995). The emitting plants synthesise isoprene in plastids (chloroplasts) from pyruvate and glyceraldehyde-3-phosphate, most probably to better resist the thermal stress (Harley et al., 1999; Kesselmeier and Staudt, 1999; Logan et al., 2000; RodriquezConcepcion and Boronat, 2002). However, many plants do not synthesise isoprene at all. Other natural sources of isoprene include sea phytoplankton (1 Tg C per year, Matsunaga et al., 2002), microbes (Wagner et al., 1999; Fall and Copley, 2000), animals and humans (Fenske and Paulson, 1999; Diskin et al., 2003). In humans, the synthesis of isoprene takes a mevalonate path that is related to the synthesis of cholesterol (Karl et al., 2001). Isoprene exhaled by humans totals roughly 4 Tg C per year. Anthropogenic sources of isoprene, which generally are poorly quantified, include combustion and evaporation of fuels (Broderick and Marnane, 2002; Borbon et al. 2003a, 2003b), cigarette smoking (Pouli et al., 2003; Baek and Jenkins, 2004), biomass burning (Andreae and Merlet, 2001), garden waste (Wilkins and Larsen, 1996) and chemical technology (Leber, 2001). Potential health risks induced the research interest in emissions of isoprene with vehicle exhaust and with cigarette smoke, as well as from facilities manufacturing natural and synthetic rubber, where concentrations of isoprene in ambient air are relatively high. Isoprene is not stored by organisms or in the environment; it encounters sinks wherever it is synthesised or transported, if only reactants are available. Endogenous isoprene may react with oxidants in the intracellular and extracellular fluids, or at walls and membranes in living organisms (Wannaz et al., 2003). The plausible reactants are oxidative species, such as free radicals or radical ions, or ozone. Expelled through the leaf stomata, or with human or animal breath, isoprene may react at the interface or in the volume of fluid films covering the body surfaces; leaves, membranes or skin. Unreacted isoprene is transported to the atmosphere and consumed in several gas-phase reactions (Atkinson and Arey, 2003, 2003a). These reactions are considered a main sink for isoprene in the environment. The possibility of heterogeneous reactions of isoprene on airborne aerosols and of multiphase reactions in atmospheric waters is discussed in this overview. Isoprene may also react in seawater, where 261 I. Barnes and K. J. Rudzinski (eds.), Environmental Simulation Chambers: Application to Atmospheric Chemical Processes, 261–277. © 2006 Springer. Printed in the Netherlands.

K. J. Rudzinski

262

it is produced, or at the sea/air interface. Up to now, only the concentration of isoprene in seawater and the fluxes of isoprene desorbing from the sea to the atmosphere have been measured, while the possibility of aqueous-phase reactions has only been suggested (Milne et al., 1995; Matsunaga et al., 2002). Little is known on the chemistry of isoprene in soils, which naturally provide opportunities for gas-phase, heterogeneous and multiphase reactions. Some soil microbes are a well-quantified sink for isoprene with recognised biochemistry (Boyd et al., 2000), while other microbes produce isoprene (Wagner et al., 1999). The overall balance of isoprene exchange with soils has been estimated at 20.4 Tg C y-1, in favour of deposition (Cleveland and Yavitt, 1997). Gas-phase transformations of isoprene The gas-phase chemistry of isoprene is beyond the scope of this overview, but a short summary is given for completeness, and because many of the gas-phase descendants of isoprene can condense, dissolve in water or react on surfaces under atmospheric conditions.

OH

•

+ O2

OH

+

NO3

•

ONO2 O3

Figure 1.

OO • OH

+ O2

OO •

ONO2

O O O

Gas-phase reactions of isoprene with atmospheric radicals and ozone (for simplicity, only one isomer of each product is shown).

Isoprene does not photolyse in the atmosphere, but reacts with hydroxyl or nitrate radicals and, successively, with molecular oxygen, or with ozone (Figure 1). A hydroxyl radical adds to either of two double bonds in the isoprene molecule to give a hydroxyalkyl radical (four isomers are possible). Hydroxyalkyl radicals react with molecular oxygen to give hydroxyalkyl peroxy radicals in seven isomeric forms. Nitrate radicals react in a similar manner to yield nitrooxyalkyl and nitrooxyalkyl peroxy radicals, each in eight isomeric forms. Ozone adds to either of double bonds of isoprene yielding two isomeric ozonides. Peroxy radicals can react with other radicals, such as NO, HO2 or RO2. A broad spectrum of possible products comprises hydroxy hydroperoxides, diols, hydroxycarbonyls, methyl vinyl ketone, methacrolein, 3-methylfuran, various nitrates (nitroxy peroxy-, hydroxy- and carbonyl-), unsaturated dinitrates, methyl glyoxal, glyoxal, glycol aldehyde, peroxyacetyl nitrate (PAN) and methacrolyl nitrate (MPAN). Ozonides decompose into Criegee radicals and formaldehyde, methyl vinyl ketone or methacrolein. The ultimate products of gas-phase transformation of isoprene in the atmosphere are carbon monoxide and carbon dioxide. Extensive summaries of the gas-phase chemistry of isoprene can be found in the Master Chemical Mechanism (Saunders et al., 2003; http://www.chem.leeds.ac.uk/Atmospheric/

Heterogeneous and Aqueous-Phase Transformations of Isoprene

263

MCM/ mcmproj.html), or in several reviews (Le Bras et al., 1997; Atkinson, 2000; Atkinson and Arey, 2003, 2003a; Fan and Zhang, 2004). Heterogeneous transformations of isoprene The research interest in heterogeneous transformations of isoprene developed recently, in response to the discovery of humic-like substances (HULIS) in atmospheric aerosols from rural and urban regions (Havers et al., 1998). HULIS are macromolecular or polymeric substances that have dark yellow to brown colour and strongly absorb UV and VIS light (Limbeck et al., 2003; Feng and Möller, 2004). They dissolve in water and contain polysaccharidic and aliphatic substructures (Havers et al., 1998), as well as mono-, di- and polycarboxylic acids (Decesari et al., 2002; Chan and Chan, 2003; Gelenceser et al., 2003; Kiss et al., 2003). Originally, no gas-phase mechanisms could explain the formation of HULIS in the atmosphere, so researchers turned to the heterogeneous transformations of isoprene as a possible explanation. Furthermore, the modellers suggested looking for heterogeneous sinks for isoprene, because the global balance of isoprene did not close in their models (von Kuhlmann et al., 2004). Limbeck et al. (2003) tested the “heterogeneous” hypothesis in a simple laboratory experiment. They passed a stream of synthetic air containing 200 – 2000 ppbv of isoprene through a quartz-fibre bed impregnated with sulphuric acid (50 – 3600 ng). The gas flow rate was 400 dm3 h–1 and the average residence time of isoprene in the bed was 10 ms. The authors varied the acidity of the quartz bed by replacing part of H2SO4 with (NH4)2SO4. In further experiments, they added ozone to the gas stream as a competitive gas-phase reactant. The yellowish products formed on the quartz bed were in part analysed by a thermogravimetric method, and in part extracted with water and, subsequently, with methanol for further analysis. The thermograms indicated the presence of low volatile products, while the water-extracts strongly absorbed the UV light. Methanol extracts were purified by solid phase extraction, from inorganic and polar organic compounds and dried to orange coloured residues. Diffusion reflectance infrared Fourier transform (DRIFT) spectra of these residues revealed the carboxylic, hydroxyl and carbonyl functionalities present. The formation rate of the products decreased with decreasing acidity of the bed, or when ozone was added to the inlet gas stream. Based on these results, Limbeck et al. (2003) suggested that the polymeric substances were formed in a heterogeneous reaction of isoprene with sulphuric acid. Moreover, they showed that the DRIFT spectra of the water-insoluble fractions of their products were extremely similar to spectra of the reference NIST dust. Czoschke et al. (2003) used two large Teflon bags (500 dm3 each) to study the formation of secondary aerosols from isoprene in the presence of ozone, at 22 – 27 oC. They filled the bags with clean, dry air containing 0.57 – 0.59 ppm of O3, and added 5 – 50 µg m–3 of seed aerosol prepared by mixing in various proportions and drying the 6.7 mM solution of (NH4)2SO4 and 5 mM solution of H2SO4. Then, they introduced gaseous isoprene or isoprenealcohol (nonanol or decanol) mixtures into the bags and measured the concentration and size distribution of particles forming, as well as the post-reaction concentrations of ozone. The authors calculated the aerosol yield as the ratio of the volume concentration of aerosol produced to the volume of reactive organic gas consumed. The latter value was estimated from the amount of ozone available to the system and the amount of OH radicals that would be produced in the isoprene-ozone reaction. Table 1 shows the initial concentrations of reactants and the calculated yields of aerosol. The amount of seed aerosol never exceeded

K. J. Rudzinski

264

0.02 to 0.5 % of the total aerosol formed. The yields were higher for acidic seed aerosol and in the presence of alcohols. Table 1. Aerosol yields in seeded ozone-isoprene-alcohol systems (Czoschke et al., 2003). Initial concentrations, ppm: Ozone Isoprene Alcohol Temperature, oC Seed aerosol Aerosol yield, %

3.7 no 24 neutral acidic 0.2 0.6

0.57-0.59 3.1 3.6 (nonanol) 27 neutral acidic 0.4 1.1

4.4 1.4 (decanol) 22 neutral acidic 0.7 1.6

The FTIR spectra showed that the aerosols formed contained hydrates, polymeric forms, acetals and hemiacetals. Several peaks in these spectra closely corresponded to peaks in FTIR spectra of aerosols collected in the Smoky Mountains (Jang et al. 2002). Czoschke et al. (2003) concluded, that in their experiments, isoprene had been oxidised to aldehydes by ozone in the gas phase, and the aldehydes had been converted to the observed products in heterogeneous, acid-catalysed reactions. However, the presented analysis of aerosol yields was very approximate, and did not exclude the possibility of heterogeneous reactions of isoprene in the discussed experiments. All the differences in reactors and techniques given, in the experiments by Limbeck et al. (2003) ozone reacted with isoprene in the gas phase, but decreased the amount of polymers formed. Thus, the heterogeneous reactions of isoprene can be a source of several components of atmospheric aerosols. The evidence is limited, however, and has to be supported by further experimental work. Other work has shown that the formation of polymeric compounds is also possible in non-acid or unseeded gas-phase experiments. Aqueous-phase transformations of isoprene Interest in the aqueous-phase reactions of isoprene first stemmed from attempts to clean the effluents from the natural rubber industry. Elkanzi and Bee Kheng (2000) studied the oxidation of isoprene in a small, stirred photoreactor (0.25 dm–3 volume, maximal lightsource intensity of 2.25 W m–2 at 254 nm and 75 mm distance). They irradiated aqueous solutions containing 1.47 mM of isoprene and 0 – 9 mM of H2O2 (pH = 6.8), for 1 to 3 hours, at 24 oC. The concentrations of isoprene and H2O2 in the reacting solutions were followed offline, using UV spectrophotometry (254 nm) and classical wet-chemical analysis methods. The rate of isoprene degradation was evaluated and best described by Equation 1: 

d >isoprene@ dt

k1>isoprene@  k 2 >isoprene@ >H2O2 @

(1)

The rate constants k1 = 3.67×10 – 4 s – 1 and k2 = 4.77×10 – 2 M – 1 s – 1 depended, in an unresolved way, on the intensity of irradiating light. To determine the reaction products, the authors extracted a few samples of the reacting solutions with ethyl acetate, acidified and dried the extracts, and analysed them with GC/MS. At short reaction times, they observed peaks corresponding to isoprene, formaldehyde and methyl vinyl ketone. In experiments with the highest initial ratio of H2O2 to isoprene, these peaks disappeared after 2 hours.

Heterogeneous and Aqueous-Phase Transformations of Isoprene

265

Pedersen and Sehested (2001) investigated the kinetics of the reaction of isoprene with ozone, being inspired by the smell of wet linen drying in the sun that was commonly attributed to ozone, but might originate from dilute products of isoprene oxidation – methacrolein and methyl vinyl ketone. These authors used a stopped-flow apparatus to mix the 47.1 µM aqueous solution of isoprene with a 6 µM solution of O3, at 5 – 30 oC and pH = 2 (the solution pHs were adjusted with HClO4, for stabilisation of ozone). They followed the extent of reaction in time by measuring the absorption of UV light by ozone, at 260 nm. Similar experiments were carried out for reactions of methacrolein or methyl vinyl ketone with ozone (initial concentrations were 59.5 µM and 57.8 µM, respectively). The reactions were first order with respect to ozone and first order with respect to the organic reactant. The rate constants and energies of activation determined by Pedersen and Sehested (2001) are shown in Table 2. Based on the low-resolution UV spectra of the reacting isoprene-ozone mixtures, these authors suggested a rapid formation of an intermediate, which could be a stabilised carbonyl oxide (Criegee radical) known from the mechanism of the gas-phase ozonolysis of isoprene. Table 2.

Second-order rate constants for reactions of isoprene, methyl vinyl ketone (MVK) and methacrolein (MACR) with ozone, determined by Pedersen and Sehested (2001).

Isoprene + ozone MACR + ozone MVK + ozone

Energy of activation

Rate constant

Reaction

M

–1

s

–1

kJ mol – 1

5 oC

30 oC

25 oC

5 – 30 oC

2.23×105 1.17×104 2.6×104

4.64×105 2.94×104 5×104

(4.2±0.2)×105 (2.4±0.1)×104 (4.4±0.2)×104

19.9±0.5 23.9±0.5 18.0±0.5

Very recently, Claeys et al. (2004, 2004a) discovered two isomers of 2-methyl-butane1,2,3,4-tetraols in aerosol samples collected over the Amazon rain forest and over the continental K-Puszta forest in Hungary. The samples also contained carboxylic acids and polysaccharidic compounds, such as malic acid, oleic acid, palmitic acid, stearic acid, arabitol, mannitol, glucose and levoglucosan. Originally, the authors suggested that tetraols were formed in a series of gas-phase reactions starting from the oxidation of isoprene by OH and O2 radicals, and just condensed to aerosols (Claeys et al., 2004). Such a mechanism required low NOx conditions, but in subsequent field campaigns, more tetraols were found at high NOx conditions (Claeys et al., 2004a). Thus, the authors hypothesised that tetraols were formed in acid-catalysed reaction of isoprene with hydrogen peroxide, presumably heterogeneously or in the aqueous-phase. They carried out simple, experiments at room temperature, in which they mixed vigorously 50 µl of isoprene, 50 µl of 50% aqueous solution of H2O2 and 25 µl of 0.1 M aqueous solution of H2SO4. The mixtures were acidic, with pH = 2. After two hours, the solutions were analysed by GC/MS, using trimethylsilyl derivatisation.. The only products found were the erythro and threo isomers of 2-methyl-butane-1,2,3,4-tetraol. The tetraols peaks were identified in GC/MS chromatograms obtained from aerosols collected in the field campaigns and processed in the same way as the laboratory samples. In addition, Claeys et al. (2004a) carried out similar experiments in which they replaced isoprene with methacrolein or methacrylic acid, the products of the gas-phase oxidation of isoprene. In both cases 2,3-

K. J. Rudzinski

266

dihydroxymethacrylic acid was produced, as well as some precipitate that has not been analysed yet.

1.0

initial concentration of isoprene, mM:

[O2] [O2]o

0.0319 0.0200 0.0040 0

0.5

(a) 0.0

0

1

2

3

4

5

-3

time × 10 , s

1.0

0.0814 0.0407 0.0235 0.0118 0.0061 0.0024 0

initial concentration of isoprene, mM:

[O2] [O2]o 0.5

(b) 0.0 0

2

4

6

8

-3

time × 10 , s

Figure 2.

Decay of oxygen during autoxidation of Na2SO3 initiated by thermal decomposition of K2S2O8 (a) or by Mn(III) (b), and inhibited by isoprene. Initial concentrations of reactants were: K2S2O8 1 mM or MnSO4 0.01 mM; Na2SO3 1.02 mM; O2 0.22-0.25 mM; for isoprene –

Our own interest in the aqueous-phase chemistry of isoprene stemmed from the research on the inhibition of autoxidation of SO2 and its possible impact on the long-range transport of this pollutant (Pasiuk-Bronikowska et al., 2000). Addition of dissolved isoprene slowed down the homogeneous autoxidation of sodium sulphite solutions initiated by thermal decomposition of K2S2O8 (Rudzinski and Pasiuk-Bronikowska, 2001) or catalysed by MnSO4 (Rudzinski et al., 2000). The experiments were performed in a stirred tank reactor operated in a batch manner, at 25 oC. The reactor contained no gas phase and was sealed to prevent any gas-liquid mass exchange. The reacting solutions were slightly alkaline, with pH = 8.0 – 8.5.

Heterogeneous and Aqueous-Phase Transformations of Isoprene

267

We followed the extent of autoxidation by measuring the concentration of dissolved oxygen versus the reaction time (Figure 2). The plots and the formulas approximating the rates of autoxidation (Eqs 2, 3) clearly show the inhibiting role played by isoprene.



d >O 2 @ k1 >O 2 @ k a1 >O 2 @ dt 1  k >isoprene @0

(for K2S2O8 experiments)

(2)



d >O 2 @ ka 0 dt

(for MnSO4 experiments)

(3)

k0 1  k c >isoprene @0

Further experiments were carried out to explain the kinetics of isoprene degradation during the autoxidation of S(IV) (Rudzinski, 2004). The reactor was equipped with a closed sampling loop for recording the UV spectra of reacting solutions in time. The spectra were decomposed to the component spectra of isoprene, Na2SO3 and K2S2O8, and the time traces of the reactant concentrations were determined (Figure 3). The results were analysed using the chain mechanism of S(IV) autoxidation to which a reaction of isoprene with SO4• – radicals was added:

CH 2 C(CH 3 )CHCH 2  SO x4 o products,

k 2,iso

(I)

The lower limit of the rate constant k2,iso for reaction (I) was estimated under a few quantitative approximations, and the upper limit required the pseudo-stationary-state approximation for SO4• – radicals and SO52 – ions (Egn 4). (7 .07 r 1 .22 ) u 10 8 M1s 1 d k 2, iso d (2 .12 r 0 .37 ) u 10 9 M1s 1

(4)

The real value of k2,iso was anticipated close to the upper limit. The errors given in (4) are just standard errors and do not reflect the errors that could originate from the assumption that isoprene reacted exclusively with sulphate radicals. Approximate analysis showed that if isoprene reacted also with sulphite radicals, the upper limit in (4) is overestimated by a few percent, and if isoprene reacted also with peroxymonosulphate radicals as well, the upper limit is overestimated by not more than 30%. The tentative mechanism of isoprene transformation accompanying the autoxidation of S(IV) was suggested, which started with the addition of a sulphate radical to a double bond in isoprene, followed by addition of oxygen and formation of a peroxy radical. The peroxy radical reacted with sulphite ion to the alkoxy radical, which in turn reacted either with oxygen or with a sulphite ion to give respectively trans-3-methyl-4-sulphoxy-2-buten-1-al or trans-3-methyl-4-sulphoxy-2-buten-1-ol (Rudzinski et al., 2002; Rudzinski, 2004).

K. J. Rudzinski

268

1.0

[iso] [iso]o 0.5

0.0 0

1

2 -3

time × 10 , s

1.0

[S(IV)] [S(IV)]o 0.5

0.0

0

1

2 -3

time × 10 , s

Figure 3. Decay of S(IV) and isoprene during autoxidation of Na2SO3 at 25 oC and pH = 8.0 – 8.5, initiated by thermal decomposition of K2S2O8: 1 mM ( ), 0.2 mM ( ) and 0.1 mM ( ), or catalysed by 0.01 mM MnSO4 ( ). Atmospheric significance of isoprene transformations Heterogeneous formation of HULIS and secondary aerosols from isoprene, as well as the aqueous-phase formation of tetraols were demonstrated only in simple laboratory experiments. Similarities in spectra and chromatograms of the laboratory samples and of the real-life aerosols collected in the field campaigns strongly support the hypothesis that the heterogeneous and multiphase chemistry influence the atmospheric transformation and global balance of isoprene, as well as formation of atmospheric aerosols. The hypothesis, however, requires further proof and a quantitative assessment, which would take into account the competition from other highly reactive organic components of the atmosphere. Such goals can

Heterogeneous and Aqueous-Phase Transformations of Isoprene

269

only be achieved through concerted efforts of analytical, atmospheric and organic chemists, as well as modellers. Pedersen and Sehested (2002) showed that the aqueous-phase reaction of isoprene with ozone was insignificant for the processing of isoprene in the atmosphere. They estimated the overall and individual lifetimes of isoprene due to reactions with ozone and the hydroxyl radical, at 25 oC and typical in-cloud conditions. The results (Table 3) indicate that clouds generally should not contribute much to the processing of isoprene in the atmosphere. Only in the aqueous phase, were the lifetimes of isoprene due to reactions with ozone and with OH radicals comparable. Similar conclusions were drawn for methyl vinyl ketone, while for methacrolein the clouds could reduce the overall atmospheric lifetime by 50 %. Table 3.

Individual and overall lifetimes of isoprene in the atmosphere, due to reactions with 0.01 pptv OH radicals and 30 ppbv O3, at 25 oC (Pedersen and Sehested, 2001).

Liquid water content Lifetime (OH), h Lifetime (O3), h Lifetime (OH + O3), h

Gas-phase 0 1.12 29.4 -

10-7 28.6 1.08

Cloud 10-6 28.6 1.08

10-5 28.6 1.08

Aqueous phase 1 1 2.05 -

Rudzinski (2004) compared the rates of the aqueous-phase reaction of isoprene with sulphate radicals against the rates of the gas- and the aqueous-phase reactions of isoprene with OH radicals, NO3 radicals and ozone. The rates were evaluated for 25 oC, typical atmospheric concentrations of reactants (Herrmann et al., 2000), and a LWC of 10-5 and 10-4. The partitioning of reactants between phases was described using Henry’s Law. The results, shown in part in Table 4, indicated that the aqueous-phase reaction of isoprene with sulphate radicals was competitive against other reactions only in the aqueous phase and at very high values of liquid water content (LWC =10–4). Table 4. Reactant X

Concentration in the gas-phase

rX,g /(rSO4,aq LWC)* (LWC = 10-4)

rX,aq /rSO4,aq * aqueous phase

OH

0.04 pptv 0.0004 pptv 0.33 pptv 0.0017 pptv 90 ppbv 30 ppbv

684.0 68.4 38.3 0.19 196 65.4

1.89 0.02 0.2 0.001 0.2 0.066

NO3 O3 *

Relative rates of isoprene reactions with OH, NO3, O3 and SO4•in the gas phase (rX,g) and in the aqueous phase (rX,aq) at 25 oC.

concentration of SO4•- radicals in the aqueous phase was 10-12 M

The analysis presented here of the atmospheric significance of the reactions has ignored several important factors such as chemical nonlinearity of the reacting systems, microphysical redistribution of reactants in real atmospheric processes, the kinetic character

K. J. Rudzinski

270

of uptake processes and the influence of chemical reactions on the solubility and the uptake of reactants. For instance, chemical reactions coupled with the mass transfer processes to and in the droplets can significantly alter the Henry’s Law partitioning of reactants based on the bulk-phase concentrations. Nonlinearities of the coupled processes can induce instabilities and oscillations already observed in simple laboratory systems (Pasiuk-Bronikowska et al., 2000a, 2003). Furthermore, the coalescence and break-up of droplets can change the concentrations of reactants inferred from the mass transfer and chemical reaction considerations. Thus, the presented conclusions may be biased and should be re-examined using a more thorough modelling procedure. There are also general reasons that justify further research on the heterogeneous and the aqueous-phase transformations of isoprene. Several reviews on atmospheric processes reflect a common awareness of a general importance of heterogeneous and multiphase reactions (Fuzzi and Ebel, 2000; Herrmann, 2003; Ravishankara, 2003). As already mentioned, von Kuhlmann et al. (2003) showed by modelling that the global balance of isoprene did not close, and called for research on heterogeneous sinks for isoprene. Moreover, the heterogeneous and the aqueous-phase chemistry of isoprene may play a role in local air quality and health problems, that occur close to the sources of isoprene – in the cities, forests, offices and habitats, transport facilities and vehicles, and in the industry. For example, isoprene or products of isoprene reaction with ozone cause eye and airways irritation (Wilkins et al., 2001; Wolkoff and Nielsen, 2001, Rohr et al., 2002). Isoprene health risks in the rubber industry have also been studied (Bird et al., 2001). The possible heterogeneous and multiphase reaction sites in living organisms, seawater and soils were referenced in the Introduction section. partial gas-phase oxidation

ozonides, peroxides

organic compounds

submicron particles heat balance

emission to atmosphere

feeding

soil, water

blue haze

growth

larger particles

plants

Figure 4.

condensation and agglomeration

deposition, rain and snow

color white => grey => => brown => black

Feedbacks in a vegetation - atmosphere system described by Went (1960).

Last but not least come the environmental feedbacks that link atmospheric chemistry, the biosphere and climate change. Isoprene is a major biogenic trace component of the atmosphere. It was more than 40 years ago when Went (1960) discussed a specific relation between plant emissions and a blue haze observed above the countryside vegetation on sunny

Heterogeneous and Aqueous-Phase Transformations of Isoprene

271

days. He postulated a cycle that started with plants and the organic compounds they emitted, that were oxidised in the atmospheric gas-phase to compounds that condensed to minute particles that caused the blue haze, and also influenced the heat balance of the ecosystem (and consequently the growth of plants and the emissions). The minute particles grew or agglomerated to larger particles that deposited with rain and snow to build up and possibly fertilise the soil, which influenced plants and the emission of organic compounds (Figure 4). This was a beautiful example of feedback relations that might involve isoprene.

? light CO2

forest growth

T

terpenoid emission

?

heterogeneous and aqueousphase reactions biodiversity

leaf physiology, membranes, damage

?

Figure 5.

aerosol formation

gas-phase reactions

SO2, NOx, NOy, O3, OH, NO3 SO2 LRT, radical scavenging

LRT NOx, NOy

Atmosphere-biosphere-climate feedbacks induced by increased temperature and contents of CO2 (details and references are given in the text).

Five years later Rasmussen and Went (1965) conducted the first systematic study of bVOC (isoprene included) contents in the air over forests, meadows and prairies at several locations in the US. Now, at the turn of the century, each year brings at least one significant review on bVOC and isoprene or isoprene alone – Kesselmeier and Staud (1999), Fuentes et al. (2000), Sharkey and Yeh (2001), Monson (2002), Penuelas and Llusia (2001, 2003), Lerdau and Gray (2003). New feedbacks involving isoprene are often discussed (Shallcross and Monks, 2000; Fuentes et al., 2001; Kulmala et al., 2003, 2004). The role of isoprene in formation of aerosols was discussed in a review on atmospheric aerosols that was published almost simultaneously with the beginning of this Workshop (Kanakidou et al., 2004).

272

K. J. Rudzinski

The biosphere-atmosphere-climate feedbacks are quite complicated. A simplified scheme in Figure 5 shows what can happen in the atmosphere and biosphere when temperature and the CO2 concentration increase. Such conditions are expected to promote the growth of plants and, consequently, the emission of terpenoids (including isoprene), which are converted to a larger amount of aerosols that shade the Earth, also by cloud formation, to reduce the temperature (Kulmala e